JP7798288B2 - Cleaning support system and cleaning support method - Google Patents

Description

Article 30, Paragraph 2 of the Patent Act is applicable. Published in the Proceedings of the 84th National Convention of the Information Processing Society of Japan, published on February 17, 2022.

Article 30, Paragraph 2 of the Patent Act is applicable. Presented at the 84th National Convention of the Information Processing Society of Japan held at the Johoku Campus of Ehime University, National University Corporation, on March 5, 2022.

The present invention relates to a cleaning support system and a cleaning support method.

Patent Document 1 discloses an information processing system that generates a performance screen showing the cleaning results performed using a cleaning tool based on the movement of the cleaning tool detected by a movement detection unit provided on the cleaning tool. This information processing system tallyes cleaning distance, cleaning time, and cleaning frequency as cleaning results.

Japanese Patent Application Laid-Open No. 2019-192167

The information processing system in Patent Document 1 merely provides the user (cleaner) with a record of cleaning performance, and does not provide any indication as to whether the cleaning is sufficient or insufficient. Meanwhile, given the current social context, there has been an increasing demand for maintaining clean living environments (e.g., homes, offices, stores) through appropriate cleaning, for example, to curb the spread of infectious diseases. In other words, there is room for improvement in providing users with suggestions for appropriate cleaning behavior by providing them with feedback on their cleaning behavior.

The objective of the present invention is to provide a cleaning assistance system and cleaning assistance method that can provide feedback to users about their cleaning behavior.

One aspect of the present invention is a cleaning request storage unit that stores cleaning requests that are requests regarding cleaning actions;
a cleaning action acquisition unit that acquires the action of the cleaning tool being used for cleaning;
a learning device that has undergone machine learning based on training data including cleaning tool movements and corresponding cleaning actions of the user to estimate the cleaning actions of the user from the cleaning tool movements acquired by the cleaning action acquisition unit;
a cleaning behavior estimation unit that estimates the cleaning behavior of the user based on the cleaning tool behavior acquired by the cleaning behavior acquisition unit using the learning device;
and a feedback unit that compares the estimated cleaning behavior of the user with the cleaning request and returns feedback to the user.

Another aspect of the present invention is to store cleaning requests, which are requests for cleaning actions;
Obtain the behavior of the cleaning tools used for cleaning,
using a learning device that has undergone machine learning to estimate the cleaning behavior of the user from the acquired cleaning tool movements based on training data including cleaning tool movements and corresponding user cleaning behaviors;
A cleaning assistance method is provided in which the estimated cleaning behavior of the user is compared with the cleaning request and feedback is returned to the user.

According to the present invention, feedback is provided to the user regarding their cleaning behavior, allowing them to be given suggestions regarding appropriate cleaning behavior. As a result, appropriate cleaning behavior by the user is supported, making it easier to maintain a clean living environment.

1 is a diagram showing a schematic configuration of a cleaning support system according to a first embodiment of the present invention; 4 is a flowchart showing the flow of cleaning support by the cleaning support system of FIG. 1 . FIG. 1 is a plan view of an exemplary room to be cleaned. FIG. 10 is a diagram showing a schematic configuration of a cleaning support system according to a second embodiment of the present invention. 5 is a flowchart showing the flow of cleaning support by the cleaning support system of FIG. 4 . FIG. 10 is a plan view of a room to be cleaned according to another example. FIG. 10 is a diagram showing a schematic configuration of a cleaning support system according to a third embodiment of the present invention. 8 is a flowchart showing the flow of cleaning support by the cleaning support system of FIG. 7 . FIG. 7 is a plan view showing the cleaning record of the room to be cleaned in FIG. 6 .

A cleaning support system and cleaning support method according to one embodiment of the present invention will now be described with reference to the accompanying drawings. Note that the following description is merely exemplary in nature and is not intended to limit the present invention, its applications, or its uses.

[First embodiment]
1 shows a schematic configuration of a cleaning support system 1 according to a first embodiment of the present invention. The cleaning support system 1 is configured to provide feedback to a user (cleaner) 90 regarding the cleaning behavior of the user 90 with respect to a room 100 to be cleaned. In this embodiment, an example will be described in which the user 90 uses a floor mop (cleaning tool) 2 to wipe and clean a floor surface 101 of the room 100 to be cleaned. The floor mop 2 has a long, rod-shaped handle 2a that is held by the user 90 and a horizontally elongated wiping part 2b attached to the lower end of the handle 2a.

The cleaning assistance system 1 includes a cleaning action acquisition unit 10 that acquires the cleaning actions of a user 90, a server 30 that evaluates the acquired cleaning actions and generates feedback, and a speaker 3 that notifies the user 90 of the feedback.

The cleaning action acquisition unit 10 is attached to the handle 2a of the floor mop 2. Specifically, the cleaning action acquisition unit 10 is located at the center of gravity of the floor mop 2. The cleaning action acquisition unit 10 has an acceleration sensor 11, a data transmission unit 12, a start switch 13, and an end switch 14.

In this embodiment, the acceleration sensor 11 is a three-axis acceleration sensor that can measure acceleration in a first direction along the longitudinal direction of the handle 2a, a second direction perpendicular to the first direction and parallel to the longitudinal direction of the wiping portion 2b, and a third direction perpendicular to the first and second directions and parallel to the lateral direction of the wiping portion 2b when the floor mop 2 is in a specific posture (for example, when the handle 2a extends parallel to the vertical direction). Note that while an acceleration sensor capable of measuring acceleration in one or two axial directions may be used as the acceleration sensor 11, a three-axis acceleration sensor is preferred for accurately capturing cleaning operations.

The data transmission unit 12 transmits the acceleration data acquired by the acceleration sensor 11 to the server 30. Communication between the data transmission unit 12 and the server 30 is carried out via wireless communication means such as Bluetooth (registered trademark) or Wi-Fi (registered trademark), or via the Internet. Alternatively, communication between the data transmission unit 12 and the server 30 may be carried out via a wired connection, and various communication connection means can be used.

The start switch 13 is operated by the user 90 when starting measurement by the acceleration sensor 11. The end switch 14 is operated by the user 90 when ending measurement by the acceleration sensor 11.

The server 30 is a well-known computer equipped with a calculation processing unit 31 (CPU), a storage unit 32, a memory, a data receiving unit 33, and a data transmitting unit 34. A learning unit 35 and an input unit 36 are connected to the server 30. Note that the learning unit 35 and/or the input unit 36 may be integrated into the server 30.

The calculation processing unit 31 has various functions configured by the implemented programs. Specifically, the calculation processing unit 31 has a cleaning request generation unit 31a, a cleaning behavior estimation unit 31b, a cleaning behavior evaluation unit 31c, and a feedback unit 31d.

The cleaning request generation unit 31a generates cleaning requests that serve as guidelines for properly cleaning the room 100 based on information about the cleaning of the room 100 input via the input unit 36. The cleaning request includes cleaning actions, including the cleaning method and cleaning time, required for each cleaning tool. For example, when using a floor mop 2 as the cleaning tool, the cleaning request generation unit 31a assigns a first cleaning action of pushing the mop straight down to wipe areas of the floor 101 of the room 100 excluding fixtures, a second cleaning action of inserting the mop into the gaps and areas underneath the fixtures, and a third cleaning action of sweeping like a broom to other areas such as around the fixtures. The generated cleaning requests are stored in the memory unit 32.

The cleaning behavior estimation unit 31b uses the learning device 35 to estimate the cleaning behavior of the user 90 based on the acceleration data received by the data receiving unit 33. Specifically, the learning device 35 classifies the cleaning behavior of the user 90 into one of the first to third cleaning behaviors. The estimated cleaning behavior of the user 90 is stored in the memory unit 32.

The cleaning behavior evaluation unit 31c evaluates cleaning behaviors from the memory unit 32 and generates cleaning records including a first actual cleaning time for the first cleaning behavior, a second actual cleaning time for the second cleaning behavior, a third actual cleaning time for the third cleaning behavior, and a total actual cleaning time obtained by adding up the first to third actual cleaning times. The generated cleaning records are stored in the memory unit 32.

The feedback unit 31d reads the cleaning request and cleaning record from the memory unit 32, evaluates the cleaning record in light of the cleaning request, and generates feedback. Specifically, it evaluates whether the cleaning record satisfies the cleaning request and/or to what extent the cleaning request is satisfied. For example, if the cleaning request requires the first cleaning action to be performed for 5 minutes or more and the estimated first cleaning record time is 4 minutes, the feedback unit 31d evaluates that the cleaning request is not satisfied and generates feedback indicating that the satisfaction level with the cleaning request is 80%.

The feedback unit 31d then transmits the feedback to the speaker 3 via the data transmission unit 34. The feedback is notified to the user via the speaker 3.

As described above, the memory unit 32 stores cleaning requests generated by the cleaning request generation unit 31a. Therefore, the memory unit 32 constitutes a cleaning request storage unit according to the present invention. Furthermore, the memory unit 32 stores cleaning behaviors by the user 90 estimated by the cleaning behavior estimation unit 31b and cleaning performance evaluated by the cleaning behavior evaluation unit 31c.

The data receiving unit 33 receives acceleration data transmitted by the data transmitting unit 12 of the cleaning operation acquisition unit 10. The received acceleration data is stored in the memory unit 32. The data transmitting unit 34 transmits the evaluation results stored in the memory unit 32 to the speaker 3.

When a large amount of input data is input, the learner 35 uses machine learning such as supervised learning, unsupervised learning, or reinforcement learning to learn the relationships between the input data and generate a trained model that represents those relationships using parameters. In other words, the learner 35 learns the features contained in the input data and generates a model that estimates results from new input. Specifically, the learner 35 is a neural network, an SVM (support vector machine), a decision tree, or a combination of these.

In this embodiment, the learning device 35 estimates cleaning behavior based on the cleaning behavior of the user 90 acquired by the cleaning behavior acquisition unit 10. Specifically, the learning device 5 is a trained model (also referred to as cleaning AI) generated by a machine learning algorithm for estimating the user's cleaning behavior from the acceleration data acquired by the cleaning behavior acquisition unit 10, based on training data including acceleration data acquired in advance by the cleaning behavior acquisition unit 10 and the corresponding cleaning behavior of the user 90.

The learning device 35 estimates the cleaning behavior of the floor mop 2, and classifies the acquired cleaning behavior based on acceleration data into one of three categories: "cleaning behavior 1" (pushing the mop straight to wipe), "cleaning behavior 2" (inserting the mop into gaps to wipe), or "cleaning behavior 3" (sweeping like a broom). Therefore, the training data used by the learning device 35 for machine learning includes cleaning behaviors 1 to 3 and the corresponding acceleration data for each.

The input unit 36 is an input device for inputting information regarding the cleaning of the room 100 to be cleaned. The information to be input includes, for example, one or more of the following: the shape and size of the room 100; the type, shape, size, placement, and presence or absence of gaps around and below each piece of furniture placed in the room 100; the location to be cleaned in the room 100 (floor, wall, etc.); and the cleaning tools to be used for cleaning. The input unit 36 can be a device that allows manual input, such as a keyboard or tablet.

Next, with reference to the flowchart shown in Figure 2, the flow of cleaning support by the cleaning support system 1 when cleaning the exemplary room 100 to be cleaned shown in Figure 3 will be described. First, the room 100 will be described. For ease of explanation, the left-right direction in Figure 3 will be referred to as the X direction, and the up-down direction will be referred to as the Y direction. More specifically, the right side will be referred to as X1, the left side as X2, the top side as Y1, and the bottom side as Y2.

Referring to Figure 3, the room 100 to be cleaned is a rectangular room that is long in the X direction and has a wooden floor 101 and is surrounded by walls 102. A door 103 is provided at the Y2 end of the wall 102 on the X2 side of the room 100. The room 100 is 16 tatami mats in size. The room 100 is a so-called living, dining, kitchen (LDK) room that includes a living room 40 located on the X1 side, a kitchen room 60 located on the X2 side, and a dining room 50 located between the living room 40 and the kitchen room 60.

The living room 40 is rectangular and extends in the Y direction, measuring approximately 6 tatami mats in size. A television stand 41 with a television 43 placed on it is located at the Y1 end of the living room 40, and a three-seater sofa 42 is located at the Y2 end. The television stand 41 and sofa 42 extend horizontally in the X direction across the living room 40, with the back surfaces 41a and 42a on the Y side abutting against the wall surface 102 of the room or separated by a small gap so small that cleaning is not necessary, while the ends 41b and 42b on the X1 side are positioned with a gap between them and the wall surface 102. The sofa 42 rests on the floor surface 101 via multiple legs (not shown), leaving a space between the sofa 42 and the floor surface 101.

The living room 40 therefore includes the following cleaning areas: a rectangular first cleaning area R1 defined between the TV stand 41 and the sofa 42 and extending in the X and Y directions; a second cleaning area R2 defined between the TV stand 41 and the wall surface 102 on the X1 side and extending in the X direction; a third cleaning area R3 defined between the sofa 42 and the wall surface 102 on the X1 side and extending in the X direction; and a fourth cleaning area R4 defined below the sofa 42.

The dining room 50 is rectangular and elongated in the Y direction, with an area of 6 tatami mats. The dining room 50 has a shelf 51 at the Y1 end, a table 52 in the center, and four chairs 53 at each of the four corners of the table 52. The back surface 51a of the shelf 51 is generally abutting the wall surface 102, while there is a gap between it and the TV stand 41. The table 52 and chairs 53 are placed on the floor surface 101 via their legs (not shown), leaving a space between them and the floor surface 101.

The dining room 50 therefore includes the following cleaning areas: a fifth cleaning area R5 defined between the shelf 51 on the X1 side and the TV stand 41; a sixth cleaning area R6 defined below the table 52; a seventh cleaning area R7 defined below each chair 53; an eighth cleaning area R8 located on the Y2 side of the table 52 and extending in the X direction from the sofa 42 to the kitchen room 60; and a ninth cleaning area R9 located on the Y1 side of the eighth cleaning area R8 and excluding the area where the shelf 51, table 52, and chairs 53 are located.

The kitchen room 60 is rectangular and elongated in the Y direction, with an area of 4 tatami mats. A kitchen counter 61 is located at the X1 end of the kitchen room 60, and a refrigerator 62 and cupboard 63 are located at the X2 end. The kitchen counter 61 extends from the Y1-side wall 102 to the Y2 side, reaching the Y1-side end of the eighth cleaning area R8. The refrigerator 62 is positioned so that its X2-side back surface 62a and Y1-side end 62b roughly abut against the wall 102. The cupboard 63 has a gap on the Y2 side of the refrigerator 62, while its X2-side back surface 63a abuts against the wall 102 or is separated from it by a small gap so small that cleaning is not necessary.

The kitchen room 60 therefore includes, as cleaning areas, a tenth cleaning area R10 extending in the Y direction between the kitchen counter 61 and the refrigerator 62 and cupboard 63, and an eleventh cleaning area R11 extending in the X direction between the refrigerator 62 and the cupboard 63. As mentioned above, the eighth cleaning area R8 extending from the dining room 50 is located on the Y2 side of the kitchen room 60.

Next, referring to Figure 2, the flow of cleaning support by the cleaning support system 1 will be described. First, cleaning location information, which is information related to cleaning the room 100 to be cleaned, is input to the server 30 via the input unit 36 (step S101). For example, the cleaning location information includes one or more of the following: the shape and size of the room 100; the type, shape, size, placement, and presence or absence of gaps around and below each piece of furniture placed in the room 100; the cleaning location in the room 100 (floor, wall, etc.); and the cleaning tools to be used for cleaning. This information may be input each time cleaning is performed, or information input during initial setup may be read from the memory unit 32.

The cleaning request generator 31a of the server 30 then generates a cleaning request based on the cleaning location information (step S102). In this embodiment, the cleaning request generator 31a divides the room 100 into first to eleventh cleaning areas R1 to R11 based on the cleaning location information, and assigns cleaning actions appropriate to each of the first to eleventh cleaning areas R1 to R11, taking into account the size and shape of each area, for a time period corresponding to the size of each area.

Specifically, the cleaning request generator 31a assigns the first cleaning action to the first cleaning area R1, the eighth cleaning area R8, and the tenth cleaning area R10, which have linearly extending portions with a predetermined width (e.g., the longitudinal width of the wiping part 2b of the floor mop 2), for a time period corresponding to the size of each area. The cleaning request generator 31a assigns the second cleaning action to the second to seventh cleaning areas R2 to R7 and the eleventh cleaning area R11, which are the areas below each fixture and the gap areas around each fixture, for a time period corresponding to the size of each area. The cleaning request generator 31a assigns the third cleaning action to the ninth cleaning area R9, which is the other area (such as the area around the fixtures), for a time period corresponding to the size of the area.

Table 1 below shows an example of the cleaning actions and cleaning times assigned to each cleaning area for room 100.

Finally, the cleaning request generation unit 31a calculates a first cleaning request time, which is the total cleaning time for the first cleaning action, a second cleaning request time, which is the total cleaning time for the second cleaning action, a third cleaning request time, which is the total cleaning time for the third cleaning action, and a total cleaning request time, which is the total of these, and generates these as cleaning requests. The memory unit 32 stores the generated cleaning requests (step S103).

Table 2 below shows example cleaning requirements for room 100.

Next, the user 90 operates the start switch 13 of the cleaning operation acquisition unit 10 to start cleaning (step S104). Once cleaning has started, the acceleration sensor 11 begins measuring acceleration (step S105), and the measured acceleration data is transmitted to the server 30 via the data transmission unit 12 (step S106).

The server 30 then receives the transmitted acceleration data via the data receiving unit 33 (step S107). The received acceleration data is processed by the server 30. Specifically, the cleaning behavior estimation unit 31b estimates whether the cleaning behavior of the user 90 corresponds to one of the first to third cleaning behaviors based on the received acceleration data using the learning device 35 (step S108). The memory unit 32 stores the estimated cleaning behavior (step S109).

From then on, steps S105 to S109 are repeatedly performed until the user 90 operates the end switch 14 of the cleaning action acquisition unit 10 to end cleaning. Therefore, the memory unit 32 stores continuous data of the estimated cleaning behavior, i.e., the cleaning behavior at each time from the start to the end of cleaning.

Next, when the user 90 operates the end switch 14 to end cleaning (step S111), the cleaning behavior evaluation unit 31c of the server 30 evaluates the cleaning behavior read from the memory unit 32 and calculates cleaning performance from the continuous data of the estimated cleaning behavior, including a first actual cleaning time, which is the total time for which the first cleaning behavior was performed, a second actual cleaning time, which is the total time for which the second cleaning behavior was performed, a third actual cleaning time, which is the total time for which the third cleaning behavior was performed, and a total actual cleaning time, which is the total of the first to third actual cleaning times (step S112). The calculated cleaning performance is stored in the memory unit 32.

Next, the feedback unit 31d reads the cleaning request and cleaning record from the memory unit 32, and compares the first to third requested cleaning times and the total requested cleaning time in the cleaning request with the first to third actual cleaning times and the total actual cleaning time in the cleaning record to evaluate the cleaning behavior of the user 90 (step S113).

For example, as shown in Table 3 below, the feedback unit 31d generates an evaluation result that evaluates the achievement rate based on each requested cleaning time and each actual cleaning time. Alternatively, each item may be represented by a single index using a weighting function.

Next, the feedback unit 31d generates feedback to be returned to the user 90 as audio data based on the evaluation results, and outputs this to the speaker 3 via the data transmission unit 34 (step S113). For example, based on the evaluation results in Table 3 above, feedback may be generated for the achievement rates of all items, or feedback may be generated only for items with achievement rates below 100%.

The speaker 3 notifies the user of the feedback by voice (step S114). As a result, the user 90 reviews their cleaning behavior based on the feedback notified by the speaker 3, and, since their cleaning behavior is supported, for example, by performing additional cleaning behaviors that have not yet reached 100% completion, it becomes easier to maintain a clean living environment.

The cleaning support system 1 according to the above embodiment provides the following effects:

(1) The cleaning support system 1 includes a memory unit 32 in which cleaning requests, which are requests regarding cleaning actions, are stored; a cleaning action acquisition unit 10 that acquires the actions of the floor mop 2 being used for cleaning; a trained learning device 35 that has performed machine learning to estimate the user's cleaning action from the actions of the floor mop 2 acquired by the cleaning action acquisition unit 10 based on teacher data including the actions of the floor mop 2 and the corresponding cleaning actions of the user 90; a cleaning action estimation unit 31b that uses the learning device 35 to estimate the cleaning action of the user 90 based on the actions of the floor mop 2 acquired by the cleaning action acquisition unit 10; and a feedback unit 31d that returns feedback to the user 90 about the estimated cleaning action of the user 90 in light of the cleaning request.
As a result, feedback on the cleaning behavior is returned to the user 90, and suggestions on appropriate cleaning behavior can be given to the user 90. As a result, the cleaning of the user 90 is supported, and it becomes easier to maintain a clean living environment.

(2) The cleaning request includes at least one or more cleaning actions and a cleaning time for each cleaning action, the cleaning action estimation unit 31b estimates the cleaning time for each cleaning action, and the feedback unit 31d further returns feedback regarding the cleaning time for each cleaning action.
As a result, feedback regarding the cleaning time for each cleaning behavior is returned to the user 90, so that more detailed suggestions regarding cleaning behavior can be given to the user 90.

(3) The cleaning action acquisition unit 10 is attached to the floor mop 2 .
As a result, the movement of the floor mop 2 is acquired directly from the floor mop 2, and therefore the movement of the floor mop 2 is acquired more accurately than when the floor mop 2 is attached to the user 90. This improves the accuracy of the estimation of the cleaning behavior by the cleaning behavior estimation unit 31b.

(4) The cleaning action acquisition unit 10 is attached to the center of gravity of the floor mop 2 .
As a result, even when the cleaning action acquisition unit 10 is attached to the floor mop 2, it is possible to suppress changes in the position of the center of gravity of the floor mop 2. Therefore, the influence of the cleaning action acquisition unit 10 attached to the floor mop 2 on the cleaning behavior of the user 90, such as the ease of handling the floor mop 2, is suppressed.

Second Embodiment
4 shows a schematic configuration of a cleaning support system 200 according to the second embodiment. The cleaning support system 200 differs from the cleaning support system 1 according to the first embodiment in that it is configured to further provide feedback to the user 90 for each cleaning area.

Specifically, the cleaning support system 200 differs from the cleaning support system 1 according to the first embodiment in that it additionally includes a cleaning position acquisition unit 210 and a beacon transmitter 220, and that the calculation processing unit 31 additionally includes a cleaning position estimation unit 31e. It also differs from the cleaning support system 1 in that the cleaning request generation unit 31a generates a cleaning request for each cleaning area, and the cleaning behavior evaluation unit 31c evaluates cleaning behavior for each cleaning area. In the following description, components that are common to the first embodiment are assigned the same reference numerals, and their description will be omitted.

The cleaning position acquisition unit 210 is attached to the handle 2a of the floor mop 2. Preferably, the attachment position of the cleaning position acquisition unit 210 is adjusted together with the cleaning operation acquisition unit 10 so that the center of gravity of the floor mop 2 does not change. The cleaning position acquisition unit 210 has a beacon receiver 211 and a data transmission unit 212. Multiple beacon transmitters 220 are installed in the room to be cleaned, and the beacon receiver 211 receives beacon signals emitted by each beacon transmitter 220 and transmits the beacon signal reception results to the server 30 via the data transmission unit 212.

In addition, while a beacon transmitter is attached to the floor mop 2, multiple beacon receivers and data transmitters may be provided in the room, and the beacon signals received by each of the multiple beacon receivers may be transmitted to the server 30 via the respective data transmitters.

Next, regarding the calculation processing unit 31, the cleaning request generation unit 31a generates a cleaning request including cleaning actions and cleaning time for each cleaning area. That is, a cleaning request is generated for each cleaning area as shown in Table 1 described in the first embodiment.

The cleaning position estimation unit 31e detects the position of the floor mop 2 based on the beacon signal received from the cleaning position acquisition unit 210 and classifies each cleaning action into a cleaning area. As a result, the cleaning performance evaluated by the cleaning action evaluation unit 31c is associated with the cleaning area and stored in the memory unit 32.

The cleaning action evaluation unit 31c evaluates the cleaning actions read from the memory unit 32 and generates cleaning records for each cleaning area, including a first actual cleaning time for the first cleaning action, a second actual cleaning time for the second cleaning action, a third actual cleaning time for the third cleaning action, and a total actual cleaning time obtained by adding up the first to third actual cleaning times. The generated cleaning records are stored in the memory unit 32.

The feedback unit 31d reads the cleaning request and cleaning record from the memory unit 32, evaluates the cleaning record in light of the cleaning request for each cleaning area, and generates feedback. Specifically, for each cleaning area, it evaluates whether the cleaning record satisfies the cleaning request and/or to what extent the cleaning request is satisfied. For example, in a certain cleaning area, if the cleaning request requests that the first cleaning action be performed for 5 minutes or more, and the estimated first cleaning record time is 4 minutes, the feedback unit 31d evaluates that the cleaning request is not satisfied and generates feedback indicating that the satisfaction level with the cleaning request is 80%.

The feedback unit 31d then transmits the feedback to the speaker 3 via the data transmission unit 34. The feedback is notified to the user via the speaker 3.

Next, with reference to the flowchart shown in Figure 5, the flow of cleaning support by the cleaning support system 200 when cleaning the exemplary room 230 to be cleaned shown in Figure 6 will be described. First, the room 230 will be described. For ease of explanation, the left-right direction in Figure 6 will be referred to as the X direction, and the up-down direction will be referred to as the Y direction. More specifically, the right side will be referred to as X1, the left side as X2, the top side as Y1, and the bottom side as Y2.

Referring to Figure 6, the room 230 to be cleaned is a roughly square room with a wooden floor 231 and surrounded by walls 232. A door 233 is provided at the Y1 end of the wall on the X2 side of the room 230. The size of the room 230 is 8 tatami mats. The room 230 is a bedroom.

In room 230, a desk 235 and chair 236 are arranged on the Y1 side, and a bed 237 is arranged on the Y2 side. The desk 235 is arranged so as to be in approximate contact with the wall surfaces 232 on the X1 and Y1 sides, and extends horizontally in the Y direction. The chair 236 is arranged on the X2 side of the desk 235. The bed 237 is arranged so as to be in approximate contact with the wall surfaces 232 on the X1 and Y2 sides, with a gap in the X direction from the desk 235, and extends horizontally in the X direction. Meanwhile, the bed 237 is spaced apart from the wall surface 232 on the X2 side. The desk 235, chair 236, and bed 237 are placed on floor surface 231 via multiple legs (not shown), and there is a space between them and floor surface 231.

The room 230 therefore includes the following cleaning areas: a first cleaning area S1, which is roughly rectangular and defined on the X2 side of the desk 235 and chair 236 and on the Y1 side of the bed 237; a second cleaning area S2, which is defined between the desk 235 and bed 237 and extends in the Y direction; a third cleaning area S3, which is roughly rectangular and defined between the bed 237 and the X2-side wall surface 232; a fourth cleaning area S4, which is defined below the desk 235; a fifth cleaning area S5, which is defined below the chair 236; and a sixth cleaning area S6, which is defined below the bed 237.

Next, referring to Figure 5, the flow of cleaning support by the cleaning support system 200 will be described. First, cleaning location information, which is information related to cleaning the room 230 to be cleaned, is input to the server 30 via the input unit 36 (S201). For example, the cleaning location information includes one or more of the following: the shape and size of the room 230; the type, shape, size, placement, presence or absence of gaps around and below each piece of furniture placed in the room 230; the cleaning location in the room 230 (floor, wall, etc.); and the cleaning tools to be used for cleaning.

Next, the cleaning request generator 31a of the server 30 generates a cleaning request based on the cleaning location information (step S202). In this embodiment, the cleaning request generator 31a divides the room 230 into first to sixth cleaning areas S1 to S6 based on the cleaning location information, and assigns cleaning actions to each of the first to sixth cleaning areas S1 to S6 for a time period according to the size of each area.

Specifically, the cleaning request generator 31a assigns the first cleaning action to the first cleaning area S1, which has a linear portion with a predetermined width (for example, the longitudinal width of the wiping part 2b of the floor mop 2), for a time period corresponding to the area's size. The cleaning request generator 31a assigns the second cleaning action to the fourth to sixth cleaning areas S4 to S6, which are the areas below the desk 235, chair 236, and bed 237, for a time period corresponding to the area's size. The cleaning request generator 31a assigns the third cleaning action to the second and third cleaning areas S2 and S3, which are the areas around each piece of furniture, for a time period corresponding to the area's size.

Table 4 below shows an example of the cleaning requests generated by the cleaning request generation unit 31a for room 230, i.e., the cleaning actions and cleaning times for each cleaning area, and the total cleaning request time calculated by adding up each cleaning time. The memory unit 32 stores the generated cleaning requests (step S203).

Next, the user 90 operates the start switch 13 of the cleaning operation acquisition unit 10 to start cleaning (step S204). Once cleaning begins, the acceleration sensor 11 starts measuring acceleration and the beacon receiver 211 starts receiving beacon signals (position measurement signals) (step S205). The measured acceleration and the beacon signal reception results are sent to the server 30 via the data transmission units 12 and 212, respectively (step S206).

The server 30 then receives the transmitted acceleration data and signal reception results via the data receiving unit 33 (step S207). The received acceleration data and signal reception results are processed by the server 30. Specifically, the cleaning behavior estimation unit 31b estimates which of the first to third cleaning behaviors the cleaning behavior of the user 90 corresponds to based on the received acceleration data using the learning device 35. Furthermore, the cleaning position estimation unit 31e estimates which of the first to sixth areas is being cleaned based on the received signal results (step S208). The memory unit 32 stores the estimated cleaning behavior and cleaning area (step S209).

From then on, steps S205 to S209 are repeatedly performed until the user 90 operates the end switch 14 of the cleaning action acquisition unit 10 to end cleaning. Therefore, the memory unit 32 stores continuous data on the estimated cleaning behavior and cleaning area, i.e., the cleaning behavior and cleaning area at each time from the start to the end of cleaning.

Next, when the user 90 operates the end switch 14 to end cleaning (step S211), the cleaning action evaluation unit 31c of the server 30 reads the cleaning action and cleaning area from the memory unit 32 and calculates, for each cleaning area, a first actual cleaning time, which is the total time for which the first cleaning action was performed, a second actual cleaning time, which is the total time for which the second cleaning action was performed, and a third actual cleaning time, which is the total time for which the third cleaning action was performed, from the continuous data of the estimated cleaning action and cleaning area.The cleaning action evaluation unit 31c also calculates a cleaning action for each cleaning area, including a total actual cleaning action time, which is the total of the first to third actual cleaning actions (step S212).The calculated cleaning actions are stored in the memory unit 32.

Next, the feedback unit 31d reads the cleaning request and cleaning record from the memory unit 32, compares the cleaning request and cleaning record for each cleaning area, and generates an evaluation result that evaluates the cleaning behavior of the user 90 (step S212).

For example, as shown in Table 5 below, the feedback unit 31d evaluates the achievement rate for the first cleaning area S1 based on the cleaning request and cleaning results. Although not explained further, the feedback unit 31d also evaluates the other cleaning areas in a similar manner.

The feedback unit 31d then generates feedback to be returned to the user 90 based on the evaluation results and outputs it to the speaker 3 via the data transmission unit 34 (step S213). For example, feedback may be generated for each cleaning area, or may be generated collectively for the room 230. The speaker 3 notifies the user of the feedback by voice (step S214).

In the above embodiment, the position of the floor mop 2 is detected using a beacon signal, but GPS may also be used. The position of the floor mop 2 may also be detected based on still and/or video images captured by an imaging device such as the camera 4 shown by the two-dot chain line in Figure 4. Alternatively, LiDAR (Light Detection and Ranging) light detection and ranging may be used. For example, the position of the floor mop 2 may be detected by irradiating light onto the floor mop 2 from a LiDAR installed in the room 230 and observing the light reflected from the floor mop 2 with the LiDAR. Detection and ranging based on Wi-Fi (registered trademark) radio wave strength (RSSI and/or CSI) may also be used. For example, a Wi-Fi transmitter and receiver may be installed in the room 230, and when the floor mop 2 is cleaning between them, the position of the floor mop 2 may be detected based on the waveform received by the receiver using AI trained on machine learning based on the position of the floor mop 2 and the waveform received by the receiver. Alternatively, for example, a Wi-Fi transmitter may be provided in the room 230 and a Wi-Fi receiver may be provided in the floor mop 2, and the position of the floor mop 2 in the room 230 may be detected based on the reception strength of the Wi-Fi signal at the floor mop 2.

The cleaning support system 200 of this embodiment further includes a cleaning position acquisition unit 210 that acquires the position of the floor mop 2, and a cleaning position estimation unit 31e that estimates the cleaning position by the user 90 based on the position of the floor mop 2 acquired by the cleaning position acquisition unit 210, and the cleaning request further includes a request regarding the cleaning area, and the feedback unit 31d further returns feedback regarding the cleaning area.
As a result, feedback regarding the cleaning area is also returned to the user 90, and suggestions regarding the cleaning area can be given to the user 90, thereby reducing the occurrence of areas that are left uncleaned.
[Third embodiment]

Figure 7 shows the schematic configuration of the cleaning support system 300 according to the third embodiment. The cleaning support system 300 differs from the cleaning support system 200 according to the second embodiment in that it projects an image onto the cleaned area.

Specifically, the cleaning support system 300 differs from the cleaning support system 200 according to the second embodiment in that it additionally includes a projector 310 and a cleaned location display unit 31f that outputs a signal to the projector 310 to cause the calculation processing unit 31 to project an image on the cleaned location. The cleaned location display unit 31f outputs a signal to the projector 310 to display a color for the cleaned location read from the memory unit 32.

The flow of cleaning support by the cleaning support system 300 will be explained with reference to the flowchart shown in Figure 8. Steps S301 to S309 and S311 to S314 are the same as steps S201 to S209 and S211 to S214 in the cleaning support system 200, and therefore will not be explained again. In step S310, the cleaned location display unit 31f reads the locations wiped by the floor mop 2 from the memory unit 32 and outputs the data to the projector 310 so that colors are projected sequentially onto the locations.

Figure 9 conceptually illustrates a state in which colors are projected from the projector 310 onto an area that has been wiped by the user 90 using the floor mop 2. As indicated by the hatching in Figure 9, images are projected sequentially (in real time) onto the first cleaning area S1 as the user 90 wipes with the floor mop 2. For example, the color projected from the projector 310 may change depending on the number of wipes, such as projecting red for the first wipe, yellow for the second wipe, and blue for the third wipe. Alternatively, instead of projecting colors from the projector 310, an image including a pattern may be projected.

The cleaning assistance system 300 according to this embodiment includes a projector that projects colors and/or images onto the cleaned area.
As a result, the user 90 is given feedback on the areas that have been cleaned, making it easy to visually identify areas that remain uncleaned. This allows the user 90 to be given a hint as to areas that have not yet been cleaned and need to be cleaned, further reducing the occurrence of uncleaned areas. Furthermore, by changing the projected color depending on the number of wipes, multiple wipes can be reliably performed even in areas that require multiple wipes, thereby supporting the cleaning behavior of the user 90.

In the above embodiments, a floor mop 2 was used as an example of a cleaning tool, and cleaning actions using the floor mop 2 were classified into first to third cleaning actions. However, other cleaning actions using the floor mop 2 may also be added. For example, when wiping in an S-shape using the floor mop 2, cleaning actions such as sweeping may also be included. The present invention can also be applied to various cleaning tools other than the floor mop 2 (e.g., handheld mops, brooms, vacuum cleaners, etc.). In such cases, cleaning actions corresponding to each cleaning tool can be set.

Furthermore, multiple cleaning actions, such as a first cleaning action and a second cleaning action, may be assigned to one cleaning area. In this case, for example, the total cleaning time achieved by the multiple cleaning actions may be evaluated in light of the cleaning request.

In addition, while the above embodiments have been described using an example in which only one type of cleaning tool (floor mop 2) is used, the present invention can also be applied to cases in which multiple cleaning tools are used for cleaning. In this case, the acceleration data (and position data) detected for each cleaning tool can be sent to server 30 along with a signal that can identify each cleaning tool, and server 30 can analyze the cleaning motion (and position) for each cleaning tool to estimate the cleaning behavior. In this case, cleaning requests can be set for each cleaning tool, and cleaning behavior evaluations and feedback can be provided for each cleaning tool.

That is, the cleaning request may include at least one or more cleaning tools and the cleaning time for each cleaning tool, and the cleaning behavior estimation unit 31b may be configured to estimate the cleaning time for each cleaning tool, and the feedback unit 31d may be configured to further return feedback regarding the cleaning time for each cleaning tool.
As a result, feedback is returned to the user 90 regarding the cleaning time for each cleaning tool, so that more detailed suggestions regarding cleaning behavior can be given to the user 90.

In addition, in each of the above embodiments, the feedback unit 31d evaluated each cleaning action in terms of cleaning time, but instead of or in addition to this, it may be configured to evaluate the quality of the cleaning. For example, acceleration data may be used to determine whether the cleaning is appropriate. For example, if the acceleration exceeds a predetermined threshold, it may be assumed that the cleaning is undesirable, as it may cause dust to fly up. In addition, if an impactful acceleration is detected, it may be assumed that the cleaning tool has collided with fixtures or the like. In this case, the feedback unit 31d may be configured to prompt the user 90 to wipe and clean at an appropriate speed, or to provide feedback to be careful not to collide with fixtures.

That is, the cleaning request includes at least one or more cleaning actions and the cleaning quality for each cleaning action, the cleaning action estimation unit 31b estimates the cleaning quality for each cleaning action, and the feedback unit 31d may further return feedback regarding the cleaning quality.
As a result, feedback on the quality of the cleaning is returned to the user 90, and suggestions for higher quality cleaning behavior can be given to the user 90.

Furthermore, in the above embodiments, the cleaning action acquisition unit 10 is attached to a cleaning tool, but this is not limited to this. For example, it may be replaced by an acceleration sensor or the like implemented in a smart device such as a smartphone worn by the user 90.

In addition, in each of the above embodiments, an acceleration sensor 11 is used as the cleaning action acquisition unit 10, but instead of or in addition to the acceleration sensor 11, a gyro sensor may be used to measure angular velocity, and the cleaning action may be estimated based on the angular velocity.

Furthermore, as shown by the two-dot chain lines in Figures 1, 4, and 7, the system may be configured to estimate cleaning behavior based on still images and/or video captured by an image capture unit such as a camera 4 instead of or in addition to the acceleration sensor 11. In this case, the learner 35 may perform machine learning from training data including angular acceleration data or still images and/or video and corresponding cleaning behaviors, and model the system to estimate cleaning behavior based on the angular acceleration data or still images and/or video.

In addition, in each of the above embodiments, the feedback is provided via the speaker 3. However, instead of or in addition to the speaker 3, the feedback may be provided by displaying an image on the screen of a smart device worn by the user 90, or by generating vibrations from the cleaning tool or the smart device. When the feedback is provided by vibration, for example, the cleaning tool or the smart device may vibrate when a cleaning request is satisfied.
As a result, feedback on cleaning can be easily given to the user 90.

In addition, in each of the above embodiments, the server 30 is configured as a computer separate from the cleaning action acquisition unit 10, but these may also be integrated. For example, the cleaning action acquisition unit 10 and the above functions of the server 30 may be implemented in a smart device worn by the user 90. Furthermore, the cleaning action acquisition unit 10 may be attached to a cleaning tool or implemented in a smart device worn by the user 90, while the above functions of the server 30 may be implemented in a smart device worn by the user 90 that is separate from the cleaning action acquisition unit 10.

In each of the above embodiments, the cleaning request generation unit 31a is configured to generate a cleaning request based on information input about the room to be cleaned, but this is not limited to this. The user 90 may also set the cleaning request.

1. Cleaning support system 2. Cleaning equipment (floor mop)
3 Speaker 4 Camera 10 Cleaning action acquisition unit 11 Acceleration sensor 12 Data transmission unit 30 Server 31 Arithmetic processing unit 31a Cleaning request generation unit 31b Cleaning action estimation unit 31c Cleaning action evaluation unit 31d Feedback unit 31e Cleaning position estimation unit 31f Cleaned location display unit 32 Memory unit 35 Learning device 36 Input unit 90 User (cleaner)
200 Cleaning support system 210 Cleaning position acquisition unit 211 Beacon receiver 220 Beacon transmitter 300 Cleaning support system 310 Projector

Claims (13)

a cleaning request storage unit in which cleaning requests, which are requests regarding cleaning actions, are stored;
a cleaning action acquisition unit that acquires the action of the cleaning tool being used for cleaning;
a learning device that has undergone machine learning based on training data including cleaning tool movements and corresponding cleaning actions of the user to estimate the cleaning actions of the user from the cleaning tool movements acquired by the cleaning action acquisition unit;
a cleaning behavior estimation unit that estimates the cleaning behavior of the user based on the cleaning tool behavior acquired by the cleaning behavior acquisition unit using the learning device;
a feedback unit that provides feedback to the user about the estimated cleaning behavior of the user in light of the cleaning request,
The cleaning request includes at least one or more cleaning actions and a cleaning time for each of the cleaning actions;
the cleaning behavior estimation unit estimates a cleaning time for each cleaning behavior;
The feedback unit returns feedback regarding cleaning time for each cleaning action. a cleaning request storage unit in which cleaning requests, which are requests regarding cleaning actions, are stored;
a cleaning action acquisition unit that acquires the action of the cleaning tool being used for cleaning;
a learning device that has undergone machine learning based on training data including cleaning tool movements and corresponding cleaning actions of the user to estimate the cleaning actions of the user from the cleaning tool movements acquired by the cleaning action acquisition unit;
a cleaning behavior estimation unit that estimates the cleaning behavior of the user based on the cleaning tool behavior acquired by the cleaning behavior acquisition unit using the learning device;
a feedback unit that provides feedback to the user about the estimated cleaning behavior of the user in light of the cleaning request,
The cleaning request includes at least one or more cleaning tools and a cleaning time for each of the cleaning tools;
the cleaning behavior estimation unit estimates a cleaning time using the cleaning tool;
The feedback unit returns feedback regarding cleaning time using the cleaning tool. a cleaning request storage unit in which cleaning requests, which are requests regarding cleaning actions, are stored;
a cleaning action acquisition unit that acquires the action of the cleaning tool being used for cleaning;
a learning device that has undergone machine learning based on training data including cleaning tool movements and corresponding cleaning actions of the user to estimate the cleaning actions of the user from the cleaning tool movements acquired by the cleaning action acquisition unit;
a cleaning behavior estimation unit that estimates the cleaning behavior of the user based on the cleaning tool behavior acquired by the cleaning behavior acquisition unit using the learning device;
a feedback unit that provides feedback to the user about the estimated cleaning behavior of the user in light of the cleaning request,
The cleaning request includes at least one or more of the cleaning actions and a cleaning quality for each of the cleaning actions;
the cleaning behavior estimation unit estimates a cleaning quality for each cleaning behavior;
The feedback unit returns feedback regarding the quality of the cleaning . a cleaning position acquisition unit that acquires the position of the cleaning tool;
a cleaning position estimation unit that estimates a cleaning position to be performed by the user based on the position of the cleaning tool acquired by the cleaning position acquisition unit,
The cleaning request further includes a request for a cleaning area;
The feedback unit further provides feedback about the cleaning area.
The cleaning support system according to any one of claims 1 to 3 . the cleaning action acquisition unit includes an acceleration sensor capable of measuring the acceleration of the cleaning tool and/or a gyro sensor capable of measuring the angular velocity of the cleaning tool ,
the cleaning action acquisition unit acquires the measured acceleration and/or angular velocity of the cleaning tool as a value representing the action of the cleaning tool;
the learning device estimates the cleaning behavior of the user based on the acceleration and/or angular velocity of the cleaning tool measured by the cleaning behavior acquisition unit ;
The cleaning support system according to any one of claims 1 to 4 . The cleaning action acquisition unit is attached to the cleaning tool.
The cleaning support system according to any one of claims 1 to 5 . The cleaning action acquisition unit is attached to the center of gravity of the cleaning tool.
The cleaning support system according to claim 6 . the cleaning operation acquisition unit includes an image capture unit capable of capturing still images and/or videos of the cleaning tool ,
the cleaning action acquisition unit acquires a still image and/or a video of the captured cleaning tool as an image representing the action of the cleaning tool;
the learning device estimates the cleaning behavior of the user based on a still image and/or a video of the cleaning tool captured by the cleaning behavior acquisition unit ;
The cleaning support system according to any one of claims 1 to 7 . the feedback unit returns the feedback by sound, image display, and/or vibration.
The cleaning support system according to any one of claims 1 to 8 . a projector that projects colors and/or images onto the cleaned area;
A cleaning support system according to claim 4 and any one of claims 5 to 9 dependent on claim 4 . storing a cleaning request, which is a request for a cleaning action;
Obtain the behavior of the cleaning tools used for cleaning,
using a learning device that has undergone machine learning to estimate the cleaning behavior of the user from the acquired cleaning tool movements based on training data including cleaning tool movements and corresponding user cleaning behaviors;
providing feedback to the user based on the estimated cleaning behavior of the user in light of the cleaning request;
The cleaning request includes at least one or more cleaning actions and a cleaning time for each of the cleaning actions;
estimating the cleaning behavior includes estimating a cleaning time for each of the cleaning behaviors;
The cleaning assistance method , wherein the feedback includes feedback regarding cleaning time for each cleaning action . storing a cleaning request, which is a request for a cleaning action;
Obtain the behavior of the cleaning tools used for cleaning,
using a learning device that has undergone machine learning to estimate the cleaning behavior of the user from the acquired cleaning tool movements based on training data including cleaning tool movements and corresponding user cleaning behaviors;
providing feedback to the user based on the estimated cleaning behavior of the user in light of the cleaning request;
The cleaning request includes at least one or more cleaning tools and a cleaning time for each of the cleaning tools;
estimating the cleaning behavior includes estimating a cleaning time using the cleaning tool;
The cleaning assistance method, wherein the feedback includes feedback regarding cleaning time using the cleaning tool. storing a cleaning request, which is a request for a cleaning action;
Obtain the behavior of the cleaning tools used for cleaning,
using a learning device that has undergone machine learning to estimate the cleaning behavior of the user from the acquired cleaning tool movements based on training data including cleaning tool movements and corresponding user cleaning behaviors;
providing feedback to the user based on the estimated cleaning behavior of the user in light of the cleaning request;
The cleaning request includes at least one or more of the cleaning actions and a cleaning quality for each of the cleaning actions;
estimating the cleaning behavior includes estimating a cleaning quality for each of the cleaning behaviors;
The cleaning assistance method, wherein the feedback includes feedback regarding the quality of the cleaning.