JP2022012820A - Self-propelled cleaner, server, self-propelled cleaner control method, and self-propelled cleaner control program - Google Patents

Abstract

To provide a self-propelled cleaner capable of performing efficient cleaning by predicting a dust quantity in a cleaning object area even when learning of a dust quantity in the cleaning object area is insufficient.SOLUTION: A self-propelled cleaner 1 comprises: a map acquisition section 24 for acquiring, as a reference map, a dust quantity map including shape data similar to an object floor surface shape of a cleaning object area being a cleaning object of the self cleaner from a map database 51 which registers a plurality of dust quantity maps including shape data 31 indicating a floor surface shape and a dust quantity associated for each coordinate of a floor surface; a dust quantity prediction section 25 for predicting a dust quantity for each coordinate of the object floor surface shape based on the object floor surface shape and the acquired reference map; and a cleaning control section 21 for controlling a cleaning operation of the self cleaner based on the predicted dust quantity.SELECTED DRAWING: Figure 1

Description

The present invention relates to a self-propelled vacuum cleaner, a server, a control method for the self-propelled vacuum cleaner, and a control program for the self-propelled vacuum cleaner.

As a conventional technique, a self-propelled vacuum cleaner provided with a dust mass position learning means for learning a position where a large amount of dust is found based on a dust amount detected during traveling is known. This self-propelled vacuum cleaner increases the suction power when it reaches the position where a large amount of dust is learned in advance, or lowers the traveling speed to improve the dust collection efficiency, and reduces the amount of dust left behind.

Japanese Unexamined Patent Publication No. 2005-204872 (published on August 4, 2005) Japanese Unexamined Patent Publication No. 2007-117146 (published on May 17, 2007)

However, the above-mentioned conventional techniques have the following problems. Specifically, it is necessary to sufficiently learn the amount of dust before it becomes possible to perform efficient cleaning, and for sufficient learning, self-propelled in the same cleaning target area. It is necessary to operate the vacuum cleaner a considerable number of times. In particular, when using a new self-propelled vacuum cleaner, when cleaning multiple areas to be cleaned with various floor shapes with one self-propelled vacuum cleaner, the amount of learning is insufficient and efficiency is high. Difficult to clean.

One aspect of the present invention realizes a self-propelled vacuum cleaner made in view of the above.

In order to solve the above-mentioned problems, the self-propelled vacuum cleaner according to one aspect of the present invention has a dust amount including shape data indicating the floor surface shape and the dust amount associated with each floor coordinate. A map acquisition unit that acquires a dust amount map containing shape data similar to the target floor shape of the cleaning target area to be cleaned by the own machine as a reference map from a map database in which multiple maps are registered, and the above. Cleaning of own machine based on the dust amount prediction unit that predicts the dust amount for each coordinate of the target floor surface shape based on the target floor surface shape and the acquired reference map, and the predicted dust amount. It is equipped with a cleaning control unit that controls the operation.

In order to solve the above-mentioned problems, the self-propelled vacuum cleaner according to one aspect of the present invention is associated with shape data indicating the floor surface shape of the cleaning target area of the own machine for each floor surface coordinate. To store the generated garbage amount map and the map generation unit that generates the garbage amount map including the garbage amount in the map database in which the garbage amount map collected from a plurality of self-propelled vacuum cleaners is registered. The map processing unit is provided with a map processing unit for processing, and the map processing unit extracts information on the number of vertices of the floor surface shape, information on the angle of each vertice, and the distance between each vertice as a feature amount, and obtains coordinates and a dust amount value. The dust amount map is processed by generating dust amount value group data in which the data size of the above is shown in a format smaller than that of the original dust amount map.

In order to solve the above-mentioned problems, the server according to one aspect of the present invention has a plurality of dust amount maps including shape data indicating the floor surface shape and the dust amount associated with each floor surface coordinate. For the registered map database and the self-propelled vacuum cleaner that requests the dust amount map, shape data similar to the target floor shape of the cleaning target area to be cleaned by the self-propelled vacuum cleaner is provided. The map providing unit includes a map providing unit that provides a included dust amount map as a reference map, and the map providing unit includes a first feature amount of the target floor surface shape and shape data of the dust amount map registered in the map database. The self-propelled type uses the dust amount map including the shape data showing the second feature amount having the smallest error from the first feature amount as the reference map by comparing with the second feature amount of the floor surface shape shown by. Provided to the vacuum cleaner, the first feature amount and the second feature amount include the number of vertices of the floor surface shape, information about the angle of each vertices, and the distance between each vertices.

In order to solve the above-mentioned problems, the control method of the self-propelled vacuum cleaner according to one aspect of the present invention includes shape data indicating the floor surface shape and the amount of dust associated with each floor surface coordinate. A map acquisition step to acquire a dust amount map containing shape data similar to the target floor shape of the cleaning target area to be cleaned by the own machine as a reference map from a map database in which multiple dust amount maps are registered. And, based on the dust amount prediction step for predicting the dust amount for each coordinate of the target floor surface shape based on the target floor surface shape and the acquired reference map, and based on the predicted dust amount, the self. Includes a cleaning control step to control the cleaning operation of the machine.

In order to solve the above-mentioned problems, the control method of the self-propelled vacuum cleaner according to one aspect of the present invention associates shape data indicating the floor surface shape of the cleaning target area of the own machine with each of the coordinates of the floor surface. A map generation step for generating a garbage amount map including the amount of garbage that has been generated, and a map database in which the generated garbage amount map is registered in the map database in which the garbage amount map collected from a plurality of self-propelled vacuum cleaners is registered. The map processing step includes a map processing step to be processed for storage, and in the map processing step, information on the number of vertices of the floor surface shape, information on the angle of each vertice, and the distance between each vertice are extracted as feature quantities, and coordinates and dust are extracted. The dust amount map is processed by generating dust amount value group data showing the correspondence with the amount value in a format having a smaller data size than the original dust amount map.

According to one aspect of the present invention, efficient cleaning can be performed.

It is a block diagram which shows the main part structure of the self-propelled vacuum cleaner which concerns on one Embodiment of this invention. It is a schematic diagram which shows the outline of the cleaning system which concerns on one Embodiment of this invention. It is a figure which shows an example of the shape data which shows the floor surface shape of the cleaning target area (hereinafter, own room) which a self-propelled vacuum cleaner targets for cleaning. It is a figure which shows an example of the data structure of the search key data and the processed shape data. It is a figure which shows some example of the processed shape data included in the processed garbage amount map registered in a map database. It is a figure which shows an example of the dust amount map in which the dust amount is plotted for each coordinate of the shape data. It is a figure which shows an example of the dust amount map (hereinafter referred to as a reference map) generated about the cleaning target area of another self-propelled vacuum cleaner whose orientation is changed according to the shape data of the own room. It is a figure which shows an example of the dust amount map generated by deforming a reference map according to the shape data of a room. It is a figure which shows the intermediate data generated by the dust amount prediction unit in the process of predicting the dust amount of the cleaning target area of own machine based on a reference map. It is a figure which shows the intermediate data generated by the dust amount prediction unit in the process of predicting the dust amount of the cleaning target area of own machine based on a reference map. It is a figure which shows an example of the garbage amount map of own room predicted based on the deformed garbage amount map. It is a sequence diagram which shows the flow of the process of predicting the amount of dust and cleaning, which is executed in a cleaning system. It is a block diagram which shows the main part structure of the other self-propelled vacuum cleaner which concerns on one Embodiment of this invention. It is a sequence diagram which shows the flow of the process of collecting a dust amount map executed in a cleaning system.

[Embodiment 1]
Hereinafter, one embodiment of the present invention will be described in detail.

(Overview of cleaning system)
FIG. 2 is a block diagram showing an outline of the cleaning system. The cleaning system 100 includes a self-propelled vacuum cleaner 1, a cloud server 2 (server), and a plurality of other self-propelled vacuum cleaners 3.

As an example, the self-propelled vacuum cleaner 1 assumes a self-propelled vacuum cleaner of a new user. It is assumed that the self-propelled vacuum cleaner 1 has never cleaned the area to be cleaned. Therefore, at this point, the self-propelled vacuum cleaner 1 has shape data of the area to be cleaned, but does not have a dust amount map.

As an example, the self-propelled vacuum cleaner 3 assumes a self-propelled vacuum cleaner of an existing user. It is assumed that the self-propelled vacuum cleaner 3 has cleaned the area to be cleaned at least once. Therefore, the self-propelled vacuum cleaner 3 has a garbage amount map learned for its own room.

The cloud server 2 is communicably connected to the self-propelled vacuum cleaner 1 and the self-propelled vacuum cleaner 3 via a communication network such as the Internet.

The cloud server 2 collects dust amount maps of the plurality of self-propelled vacuum cleaners 3, and responds to the request of the self-propelled vacuum cleaner 1 to cover the floor surface in the cleaning target area of the self-propelled vacuum cleaner 1. Provides a garbage map with similar shapes.

The shape data is data showing the floor surface shape of the cleaning target area (own room) that the self-propelled vacuum cleaner targets for cleaning. The shape data includes, for example, coordinate information in a two-dimensional coordinate system of each vertex of the floor surface shape, and information such as orientation (direction).

The dust amount map is data in which the dust amount at the position of the cleaning target area corresponding to the coordinates of the shape data is plotted. Of the dust amount maps, the dust amount map of the other self-propelled vacuum cleaner 3 provided to the self-propelled vacuum cleaner 1 is predicted based on the reference map, and the self-propelled vacuum cleaner 1 predicts based on the reference map. The obtained garbage map in the room is called a prediction map.

The self-propelled vacuum cleaner 1 transmits the shape data of the cleaning target area of the own machine generated by the own machine to the cloud server 2 and requests a reference map. The cloud server 2 extracts a dust amount map including a floor surface shape similar to the floor surface shape indicated by the shape data of the request source, and provides the self-propelled vacuum cleaner 1 as a reference map. As an example, the cloud server 2 may extract a dust amount map including a floor surface shape most similar to the floor surface shape indicated by the shape data. It should be noted that the most similar floor surface shape may mean that the shape matches the floor surface shape indicated by the shape data. That is, the cloud server 2 can extract a dust amount map including a floor surface shape that matches the floor surface shape indicated by the shape data.

As a result, the self-propelled vacuum cleaner 1 can predict the amount of dust in its own room and generate a prediction map by referring to the dust amount map having a floor surface shape similar to (similar to) its own room. Then, the self-propelled vacuum cleaner 1 can execute a cleaning operation according to the amount of dust based on the prediction map, and as a result, efficient cleaning can be performed even if the learning about the own room is insufficient. Can be carried out.

(Hardware configuration of self-propelled vacuum cleaner)
FIG. 1 is a block diagram showing a configuration of a main part of the self-propelled vacuum cleaner 1. The block diagram shown in FIG. 1 only partially shows an example of the configuration of a self-propelled vacuum cleaner. The self-propelled vacuum cleaner 1 may include other members (not shown) included in a general self-propelled vacuum cleaner.

As shown in FIG. 1, the self-propelled vacuum cleaner 1 has, as an example, a control unit 10, a communication unit 11, an operation panel 12, a storage unit 13, a traveling unit 14, a brush unit 15, a suction unit 16, and a collision detection unit 17. , And a dust detection unit 18.

The control unit 10 sets each block in the self-propelled vacuum cleaner 1 based on the programs and data stored in the storage unit 13, the data input from the operation panel 12, the programs and data received from the external device, and the like. Control in a unified manner. The control unit 10 may be configured by, for example, a CPU (central processing unit) or an arithmetic processing unit such as a dedicated processor.

The communication unit 11 transmits / receives data to / from an external device such as a cloud server 2 via a network. For example, the communication unit 11 transmits the shape data generated by the self-propelled vacuum cleaner 1 to the cloud server 2. Further, the communication unit 11 receives the reference map provided by the cloud server 2.

The communication unit 11 may receive the feedback signal from the charging stand. Upon receiving the return signal, the self-propelled vacuum cleaner 1 returns to the installation location of the charging stand installed in any of the cleaning target areas and docks with the charging stand. As a result, a battery (not shown) is charged.

The operation panel 12 functions as a user interface. The operation panel 12 includes, for example, an operation switch that receives various instructions of the user and a display that displays various information presented to the user. The operation panel 12 may be provided as a touch panel. Further, the operation panel 12 may be provided with a display LED.

The operation panel 12 is filled with, for example, a button for instructing the start and stop of cleaning, a timer set button for making an operation reservation for the self-propelled vacuum cleaner 1, a time display area for displaying the current time and the reserved time, and a dust collecting container. It may include a garbage dump lamp or the like to inform that the time has become. Further, the operation panel 12 may include an area for displaying a battery mark indicating the charge amount of the battery. These are merely examples, and the operation panel 12 may include various input / output units for realizing other functions.

The storage unit 13 stores various data used by the control unit 10, and includes a ROM (read only memory), a RAM (Random Access Memory), and the like. In the present embodiment, the storage unit 13 non-volatilely stores the shape data 31, the reference map 32, and the prediction map 33, as an example. In addition, the storage unit 13 may store various programs (not shown) executed by the control unit 10.

The traveling unit 14 is a mechanism for self-propelling the self-propelled vacuum cleaner 1. The traveling unit 14 is composed of, for example, a traveling drive unit including a motor driver, a drive wheel motor, and the like, and drive wheels. The traveling drive unit determines the rotation direction, rotation angle, and the like based on the control signal from the cleaning control unit 21 of the control unit 10 to drive the drive wheels.

The brush portion 15 is a mechanism for collecting dust on the floor surface at the suction port. The brush unit 15 includes, for example, a rotary brush drive unit, a rotary brush, a side brush drive unit, a side brush, and the like. The rotary brush drive unit includes a motor driver, a rotary brush motor, and the like, and determines the number of rotations and the like based on a control signal from the cleaning control unit 21 to drive the rotary brush. The side brush drive unit includes a motor driver, a side brush motor, and the like, and determines the rotation speed and the like based on the control signal from the cleaning control unit 21 to drive the side brush.

The suction unit 16 is a mechanism for taking in dust on the floor surface from the suction port and storing it in the dust collecting container. The suction unit 16 includes a motor driver, a fan motor, and the like, determines the rotation speed and the like according to a control signal from the cleaning control unit 21, and changes the air volume generated in the vicinity of the suction port, that is, the suction force. Keep it constant.

The collision detection unit 17 is a contact sensor that detects that a bumper (not shown) provided in the main body of the self-propelled vacuum cleaner 1 collides with an object such as a wall or an obstacle.

The dust detection unit 18 is a sensor that measures the amount of dust taken in from the suction port, that is, the amount of dust, and the sensor is realized by, for example, an infrared sensor.

In addition to the collision detection unit 17 and the dust detection unit 18, the self-propelled vacuum cleaner 1 includes various sensors such as a distance measuring unit and a step detecting unit for measuring the floor surface shape of the cleaning target area. You may. The self-propelled vacuum cleaner 1 can measure the distance between the self-propelled vacuum cleaner 1 and an object such as an obstacle or a wall by providing a distance measuring unit. By measuring the distance between the own machine and the object, the possibility that the own machine collides with the object can be reduced. For example, since the self-propelled vacuum cleaner 1 is provided with a distance measuring unit, it can move along the wall without colliding with the wall, and shape data can be efficiently collected. The ranging unit may be composed of, for example, an ultrasonic sensor, an infrared sensor, a laser sensor, or the like.

(Software configuration of self-propelled vacuum cleaner 1)
As an example, the control unit 10 of the self-propelled vacuum cleaner 1 includes a cleaning control unit 21, a sensor control unit 22, a shape data generation unit 23, a map acquisition unit 24, and a dust amount prediction unit 25.

Each unit of the control unit 10 described above has a CPU (central processing unit) stored in a storage device (storage unit 13) realized by a ROM (read only memory), an NVRAM (non-Volatile random access memory), or the like. This can be achieved by reading the program into RAM (random access memory) and executing it.

The cleaning control unit 21 controls the cleaning operation of the self-propelled vacuum cleaner 1. The cleaning operation of the self-propelled vacuum cleaner 1 includes traveling, brush rotation, and suction operation.
The cleaning control unit 21 sends a control signal to each mechanism for cleaning, that is, the traveling unit 14, the brush unit 15, and the suction unit 16, and causes each mechanism to perform self-propelled cleaning.

The cleaning control unit 21 may store the operation history of the traveling unit 14, the brush unit 15, and the suction unit 16 in the storage unit 13.

The sensor control unit 22 acquires sensor values measured by various sensors such as the collision detection unit 17 and the dust detection unit 18, the self-propelled vacuum cleaner 1, performs necessary processing, and stores the sensor values in the storage unit 13. ..

The shape data generation unit 23 generates shape data indicating the floor surface shape of the cleaning target area (own room). The shape data generation unit 23 may generate shape data based on information regarding the floor surface shape of the own room previously input by the user. Alternatively, the shape data generation unit 23 may generate shape data of its own room based on the above-mentioned operation history and sensor values acquired by one or more runs. Alternatively, the shape data generation unit 23 may automatically generate the shape data based on the operation history and the sensor value, and then accept the editing operation by the user to complete the shape data. The shape data generation unit 23 stores the completed shape data 31 in the storage unit 13.

The map acquisition unit 24 transmits the search key data to the cloud server 2 and requests a reference map. The map acquisition unit 24 acquires the reference map searched by the cloud server 2 based on the search key data from the cloud server 2. The search key data is information indicating the shape of the area to be cleaned with a data structure simpler than that of the shape data 31 and a smaller data size. The search key data may be information indicating the feature amount of the shape of the cleaning target area extracted from the shape data 31. By using the search key data, the cloud server 2 creates a dust amount map having a shape similar to the cleaning target area of the self-propelled vacuum cleaner 1, and a huge amount of dust map registered in the map database 51. It is possible to search efficiently from among. The map acquisition unit 24 processes the shape data 31 of its own room generated by the shape data generation unit 23 into search key data by a procedure described later.

The map acquisition unit 24 stores the received reference map 32 in the storage unit 13. The map acquisition unit 24 may order the reference map periodically, for example, at a fixed time of the day, or may order the reference map each time immediately before the start of cleaning.

The garbage amount prediction unit 25 predicts the amount of garbage in the room based on the reference map. As an example, the dust amount prediction unit 25 processes a reference map so as to match the shape data of the room, and generates a predicted dust amount map of the room as a prediction map. The dust amount prediction unit 25 stores the generated prediction map 33 in the storage unit 13.

As a result, the cleaning control unit 21 controls the traveling unit 14, the brush unit 15, and the suction unit 16 based on the prediction map 33 generated by the dust amount prediction unit 25 to efficiently perform cleaning. Can be done. For example, the cleaning control unit 21 may increase the suction force or secure a long suction time at a position indicated by a large amount of dust in the prediction map 33 to improve the dust collection efficiency at that position. can.

(Cloud server configuration)
As shown in FIG. 1, the cloud server 2 includes at least a control unit 40 and a map database 51 as an example.

The control unit 40 controls each block in the cloud server 2 based on the programs and data stored in the storage unit 13, the data input from the operator of the cloud server 2, the programs and data received from the external device, and the like. To control. The control unit 40 may be configured by, for example, a CPU (central processing unit) or an arithmetic processing unit such as a dedicated processor.

The map database 51 is a database constructed in a storage device (not shown) of the cloud server 2, and stores a large number of dust amount maps uploaded from the self-propelled vacuum cleaner 3.

The control unit 40 includes a map providing unit 41 and a map registration unit 42 as an example. Each of these parts is a RAM (random) program in which the CPU (central processing unit) is stored in a storage device (not shown) realized by a ROM (read only memory), NVRAM (non-Volatile random access memory), or the like. It can be realized by reading it into access memory) and executing it.

The map providing unit 41 is most similar to the floor surface shape of the cleaning target area of the self-propelled vacuum cleaner 1 from the map database 51 based on the shape data transmitted from the self-propelled vacuum cleaner 1. Search for a garbage map (similar). Then, the map providing unit 41 provides the self-propelled vacuum cleaner 1 of the requesting source with the dust amount map having the most similar (most similar) floor surface shape hit in the search as a reference map.

The map registration unit 42 collects the dust amount map transmitted from the self-propelled vacuum cleaner 3 and registers it in the map database 51. The dust amount map registered by the map registration unit 42 may be processed by the self-propelled vacuum cleaner 3 as necessary in order for the map providing unit 41 to efficiently search for a similar reference map. ..

In one preferred embodiment, the garbage map registered in the map database 51 may be a processed garbage map having a data structure suitable for searching for garbage maps having similar shapes. The processed garbage amount map is information showing the feature amount extracted from the garbage amount map, and is information having a simple data structure and a small data size as compared with the garbage amount map. The processed dust amount map includes processed shape data and dust amount value group data.

The processed shape data is information indicating the feature amount of the floor surface shape indicated by the shape data, and is generated by processing the shape data included in the dust amount map. The processed shape data is generated for the purpose of facilitating the search for similar shapes in the cloud server 2. Therefore, it has the same data structure as the above-mentioned search key data used as a search key. The data structures of the search key data and the processed shape data will be described in detail later.

The dust amount value group data is a set of dust amount values that summarizes the dust amount values for each coordinate of the shape data. The dust amount value group data is data showing the correspondence between the coordinates and the dust amount value at the coordinates in a text format. The dust amount value group data may be text format data in which each item (X coordinate, Y coordinate, dust amount value) is separated by a comma, for example, CSV (Comma Separated Values) data. .. As a result, the data size of the processed garbage amount map is significantly reduced as compared with the original garbage amount map, the storage capacity of the map database 51 is saved, and a large amount of processed garbage amount map is registered in the map database 51. Allows you to.

When the map providing unit 41 receives the search key data from the self-propelled vacuum cleaner 1, the map providing unit 41 compares the search key data with the processed shape data of each processed dust amount map registered in the map database 51. Specifically, the map providing unit 41 has the floor surface shape (target floor surface shape) of the cleaning target area indicated by the search key data and the processed shape data of the processed dust amount map registered in the map database 51. Compare with the shown floor shape (reference floor shape). The map providing unit 41 searches for a processed dust amount map including the feature amount most similar to the feature amount of the target floor surface shape, and provides the processed dust amount map to the self-propelled vacuum cleaner 1 as a reference map. ..

The map providing unit 41 compares the feature amount of the target floor surface shape (first feature amount) shown in the search key data with the feature amount of the reference floor surface shape shown in the processed shape data (second feature amount). By doing so, the search can be performed with a lower load than when comparing the shape data with each other.

(Shape data)
FIG. 3 is a diagram showing an example of shape data. As an example, the shape data includes coordinate information of each vertex of the floor surface shape in a two-dimensional coordinate system and information on the direction (direction). Further, the shape data may include information such as the linear distance of the outer circumference of the floor surface shape and the distance between the vertices of each vertex. Further, the shape data may include information indicating the installation location of the charging stand.

In the following description, it is assumed that the illustrated shape data indicates the shape data generated by the shape data generation unit 23 of the self-propelled vacuum cleaner 1.

(Processing of shape data)
FIG. 4 is a diagram showing an example of a data structure of search key data processed from shape data. FIG. 4 further shows, in order to explain an example of the procedure until the search key data 313 is processed, the shape data 31 before processing and the data structure of the number of vertices 311 generated in the processing process. , The data structure of the angle type definition table 312 referenced for machining.

The map acquisition unit 24 processes the shape data 31 to generate the search key data 313 in order to reduce the amount of data sent / received to / from the cloud server 2 and to facilitate the search on the cloud server 2. The search key data 313 shown in FIG. 4 is shown in a table format as an example, but the search key data 313 can be in any data format, and may be, for example, a matrix or a list.

When the map acquisition unit 24 reads the shape data 31 from the storage unit 13, it detects the vertices of the shape indicated by the shape data 31. In the example shown in FIG. 4, a total of 6 vertices 1 to 6 are detected. The map acquisition unit 24 substitutes 6 for the number of vertices N and generates the number of vertices 311 as one of the feature quantities of the shape of the area to be cleaned.

Further, the map acquisition unit 24 assigns a vertex number to each of the detected vertices. The numbering rule of the vertex number is not particularly limited, but it is necessary to be unified with the numbering rule of the vertex number in the processed shape data registered in the map database 51. In the present embodiment, as an example, vertex numbers may be assigned according to the following numbering rule. First, the map acquisition unit 24 uses the upper left vertex of the shape data 31 as the origin. Then, the map acquisition unit 24 assigns a serial number to the vertices that first reach from the origin in a clockwise direction around the outer circumference of the shape data 31, and then to the vertices in the order of reaching clockwise. The map acquisition unit 24 assigns the last number, that is, the N number having 311 vertices, to the origin to be reached last.

Further, the map acquisition unit 24 defines the sides L1 to LN between the vertices and acquires the distance between the vertices. The edge LN refers to a line segment from the vertex of (N-1) to the vertex of N. The map acquisition unit 24 acquires the length of the side LN, that is, the distance from the vertex (N-1) to the vertex N, and dN (m) for each side. In the specific example shown in FIG. 4, the map acquisition unit 24 has the distance between the vertices from the side L1 to the side L6, in that order {d1 (m), d2 (m), d3 (m), d4 (m), d5 (m). ), D6 (m)}.

Further, the map acquisition unit 24 acquires the internal angle of each vertex. The search key data includes information regarding the angle of the internal angle of the floor surface shape. In the present embodiment, as an example, the map acquisition unit 24 classifies each vertex into several angle types according to the value of the internal angle instead of the value of the internal angle of each vertex, and uses the angle type as information regarding the angle. get. As a result, the search key data can be further simplified as compared with the case where the value of the internal angle of each vertex is held. In the example shown in FIG. 4, the map acquisition unit 24 refers to the angle type definition table 312 stored in the storage unit 13 in advance, and based on the value of the internal angle of each vertex, each vertex has three types of angle types. Classify either.

In the specific example shown in FIG. 4, the map acquisition unit 24 determines the angle types from the apex 1 to the apex 6 as {A, B, A, A, A, A} in order.

When the map acquisition unit 24 acquires the angle type of the internal angle of each vertex and the distance between the vertices in the shape of the cleaning target area from the shape data 31, the map acquisition unit 24 generates search key data 313 including these. The map acquisition unit 24 can send the generated search key data 313 to the cloud server 2 and request the cloud server 2 for a garbage amount map to be referred to.

The cloud server 2 can efficiently search for a dust amount map having a shape close to the cleaning target area of the self-propelled vacuum cleaner 1 by searching the map database 51 using the search key data 313 as a key. Next, an example of the search procedure on the cloud server 2 will be described.

(Search reference map)
FIG. 5 is a diagram showing some examples of processed shape data included in a record (processed dust amount map) registered in the map database 51. Hereinafter, an example of a reference map search procedure carried out by the map providing unit 41 will be described with reference to FIG. Here, it is assumed that the map providing unit 41 has received the search key data 313 shown in FIG. 4 from the map acquisition unit 24 of the self-propelled vacuum cleaner 1.

(1) Vertex number filtering The map providing unit 41 first extracts a record of processed shape data having the same number of vertices N as the number of vertices N indicated by the search key data 313. As an example, since the search key data 313 indicates the number of vertices N = 6, the map providing unit 41 extracts a record of the processed shape data indicating the number of vertices N = 6. In the example shown in FIG. 5, the fifth processed shape data 515 is screened out because the number of vertices N does not match, and the first processed shape data 511 to the fourth processed shape data 514 having the same number of vertices N are used. The records to have are extracted.

(2) Angle type filtering The map providing unit 41 next extracts records in which the angle types of each vertex are arranged in the order of the vertex numbers assigned according to the above-mentioned numbering rule. .. Specifically, the sequence of the angle types of the search key data 313 shown in FIG. 4 is clockwise {A, B, A, A, A, A}. The map providing unit 41 extracts a record having processed shape data whose angle type arrangement matches this. For example, the map providing unit 41 extracts a record having the first processed shape data 511 in which the angle type arrangement is the same {A, B, A, A, A, A} as the search key data 313.

It should be noted that extracting only the records in which the vertex numbers and the arrangement of the angle types completely match as described above has the following inconveniences. That is, the inconvenience that records having different directions of the floor surface shape or records having different origin definition methods in the floor surface shape are screened out even though the floor surface shapes are similar or similar. There is. In order to avoid such inconvenience, the map providing unit 41 repeatedly executes the above-mentioned angle type filtering process for the remaining 5 patterns obtained by sequentially replacing the heads of the angle types of the search key data 313. ..

In the example of the search key data 313, the arrangement of the angle types is the following five patterns in addition to the above-mentioned first pattern {A, B, A, A, A, A} starting with the angle type of vertex number 1. Can be considered. That is, the second pattern {B, A, A, A, A, A} with the angle type of vertex number 2 at the beginning and the third pattern {A, A, A, with the angle type of vertex number 3 at the beginning. A, A, B}, a fourth pattern {A, A, A, A, B, A} headed by the angle type of vertex number 4, and a fifth pattern {A, A, A, A, B, A} headed by the angle type of vertex number 5. There are A, A, A, B, A, A} and a sixth pattern {A, A, B, A, A, A} headed by the angle type of vertex number 6 (origin).

The map providing unit 41 extracts records having the same angle type arrangement for each of these patterns. As a result, not only the record of the first processed shape data 511 but also the record of the second processed shape data 512 that matches the sixth pattern of the search key data 313 also matches the third pattern of the search key data 313. 3 The records of the processed shape data 513 are also extracted as records having similar shapes without omission.

By adopting such a search algorithm, it is possible to absorb the difference in the orientation of the area to be cleaned and the difference in the way of taking the origin, and extract records having similar shapes without omission.

In the above-mentioned angle type filtering process, it is assumed that a plurality of records having the same angle type arrangement are found for one pattern of the angle type arrangement (any of the above-mentioned first to sixth patterns). .. In this case, the map providing unit 41 further extracts one record having the processed shape data having the smallest error in the distance between vertices.

Conversely, in the angle type filtering process, it is assumed that no record with the same angle type arrangement can be found for the angle type arrangement pattern. In this case, the map providing unit 41 extracts records in which the arrangement of angles matches as many as possible for the pattern.

In this way, one candidate record is selected for each pattern. In the example of the search key data 313, since there are 6 patterns of the angle type arrangement, 6 candidate records are selected.

(3) Reference map determination The map providing unit 41 compares the processed shape data of the selected candidate record with the search key data 313. Then, a record (processed dust amount map) having the processed shape data having the closest angle type arrangement and the smallest error in the distance between the vertices is determined as a reference map to be provided to the self-propelled vacuum cleaner 1. do.

As described above, the map providing unit 41 can efficiently search the reference map from the map database 51 with low load processing as compared with the case of comparing the shape data 31 with each other.

(Garbage amount map)
FIG. 6 is a diagram showing an example of a dust amount map. In the dust amount map, the dust amount at the position of the cleaning target area corresponding to the coordinates of the shape data is associated with each coordinate. The associated dust amount may be represented in multiple stages, as shown in the illustrated example, may be represented by a binary value of high (1) or low (0), or the measured dust amount. It may be expressed by the numerical value of.

In the following description, the illustrated dust amount map is assumed to be similar to the shape data shown in FIG. 3, and indicates a reference map provided to the self-propelled vacuum cleaner 1 by the map providing unit 41 of the cloud server 2. do.

In a more preferred embodiment, the reference map provided to the self-propelled vacuum cleaner 1 by the map providing unit 41 is a processed garbage amount map searched from the map database 51. Therefore, when the map acquisition unit 24 receives the reference map from the cloud server 2, the above-mentioned processed garbage amount map is based on the processed shape data included in the reference map and the garbage amount value group data. The reference map may be restored to the garbage amount map shown in FIG. The garbage value group data is data in which the correspondence between the coordinates and the garbage value is shown in a text format. The map acquisition unit 24 plots dust amount values at each coordinate in the XY coordinate system showing the floor surface shape developed from the processed shape data, as shown by the dust amount value group data described above, thereby producing dust amount. You can restore the map.

(Processing of reference map)
FIG. 7 is a diagram showing an example of a reference map whose orientation is changed according to the shape data of the own room shown in FIG.

For example, the dust amount prediction unit 25 compares the shape data shown in FIG. 3 with the reference map shown in FIG. 6, determines the rotation angle of the reference map with the least error, and rotates the reference map. Specifically, the dust amount prediction unit 25 rotates the reference map shown in FIG. 6 by 90 degrees counterclockwise so as to match the shape data shown in FIG. 3 as much as possible, and the reference after rotation shown in FIG. Get the map.

FIG. 8 is a diagram showing an example of a dust amount map in which the reference map shown in FIG. 6 is deformed according to the shape data of the own room shown in FIG. In the following, the deformed dust amount map will be referred to as an intermediate map 321.

For example, when the dust amount prediction unit 25 compares the shape data shown in FIG. 3 with the reference map after rotation shown in FIG. 7, it is assumed that the floor surface shapes of both are similar to each other. In this case, the dust amount prediction unit 25 uses the rotated reference map as its floor shape so that the area or outer peripheral length of the rotated reference map matches the area of the shape data or the outer peripheral length. Enlarge or reduce while maintaining. That is, the dust amount prediction unit 25 maintains the aspect ratio of the floor shape of the reference map so that the size of the floor shape of the reference map matches the size of the floor shape of the room. Determine the scaling ratio to scale the whole.

Alternatively, it is assumed that the dust amount prediction unit 25 finds that the floor surface shapes of the two have a dissimilar relationship. In this case, the dust amount prediction unit 25 enlarges or reduces each side of the rotated reference map so that the floor shape shown by the rotated reference map matches the floor shape shown by the shape data. At this time, the dust amount prediction unit 25 may determine the scaling ratio for each side forming the floor surface shape of the reference map.

For example, the dust amount prediction unit 25 determines each side of the reference map so that the length of each side of the rotated reference map shown in FIG. 7 is the same as the length of each corresponding side in the shape data shown in FIG. Shrink or zoom in. In the illustrated example, the dust amount prediction unit 25 enlarges the first side (s1) by 1.5 times in the direction of the X axis and the fourth side (s4) by 1.2 times in the direction of the Y axis. .. Further, the dust amount prediction unit 25 reduces the second side (s2) by 0.66 times in the Y-axis direction and the third side (s3) by 0.75 times in the X-axis direction. In this way, the intermediate map 321 is obtained. The intermediate map 321 is used by the dust amount prediction unit 25 to generate the prediction map 33 according to the shape data 31 of the own room.

In another example, the dust amount predictor 25 determines each of the reference maps so that each vertex of the rotated reference map shown in FIG. 7 is located at the same coordinates as the corresponding vertex in the shape data shown in FIG. You may move the vertices.

(Forecasting the amount of garbage)
9 and 10 are diagrams showing intermediate data generated by the dust amount prediction unit 25 in the process of predicting the dust amount in the cleaning target area of the own machine based on the reference map. Hereinafter, an example of a procedure for generating a dust amount prediction map in the own room, which is carried out by the dust amount prediction unit 25, will be described with reference to FIGS. 9 and 10. In the present embodiment, as an example, the dust amount prediction unit 25 arranges the dust amount value at the coordinates of the intermediate map 321 according to the scaling ratio when the reference map 32 is scaled according to the shape data 31 of the room. Relocate the distribution of. In particular, in the present embodiment, as an example, the dust amount prediction unit 25 divides the distribution region of the dust amount value of the intermediate map 321 into an X component and a Y component, and then scales the distribution region to obtain expansion / contraction parts of the distribution region. .. Finally, the scaling parts of the intermediate map 321 obtained by scaling for each component are combined to generate a prediction map 33 in which the dust amount value is arranged according to the shape of the own room.

(1) Adjustment of dust amount in the X-axis direction The dust amount prediction unit 25 adjusts the dust amount of the X component of the intermediate map 321 based on the scaling factor in the X-axis direction on each side of the reference map 32. In the present embodiment, the distribution of the dust amount value in the region sandwiched between the side expanded and contracted in the X-axis direction and the side facing the side is adjusted.

In the example shown in FIG. 9, in the intermediate map 321 the side s1 is enlarged 1.5 times in the X-axis direction. Therefore, the dust amount prediction unit 25 enlarges the region sandwiched between the side s1 and the opposite side ss1 by 1.5 times in the X-axis direction to generate the X component expansion part 322. Further, the side s3 is reduced to 0.75 times in the X-axis direction. Therefore, the dust amount prediction unit 25 reduces the region sandwiched between the side s3 and the opposite side ss3 by 0.75 times in the X-axis direction to generate the X component reduction part 323.

As an example, when the dust amount prediction unit 25 expands the dust amount value in the X-axis direction, a new element is added between two adjacent elements in the X-axis direction, and the above-mentioned two adjacent elements are added thereto. Interpolate the mean value of. For example, the interpolated dust amount value may be a value obtained by rounding off the first decimal point of the average value.

As an example, when reducing the dust amount value in the X-axis direction, the dust amount prediction unit 25 replaces two adjacent elements in the X-axis direction with one element, and averages the above-mentioned two adjacent elements there. Assign a value. For example, the assigned garbage amount value may be a value obtained by rounding off the first decimal point of the average value.

The dust amount prediction unit 25 combines the X component expansion part 322 and the X component reduction part 323 to generate an intermediate map with the X component expansion / contraction completed. In the following, this intermediate map will be referred to as an X component scaled map 331.

In addition, when the X component expansion part 322 and the X component reduction part 323 are combined, the squares (coordinates) may overlap. In this case, for example, the dust amount prediction unit 25 may associate the average value of the dust amount value in the X component expanding part 322 and the dust amount value in the X component reducing part 323 of the overlapping squares with the square. ..

(2) Adjustment of dust amount in the Y-axis direction Next, the dust amount prediction unit 25 determines the dust amount of the Y component of the X component scaled map 331 based on the scaling ratio in the Y-axis direction on each side of the reference map 32. To adjust. In the present embodiment, the distribution of the dust amount value in the region sandwiched between the side expanded and contracted in the Y-axis direction and the side facing the side is adjusted.

In the example shown in FIG. 10, in the X component scaled map 331, the side s4 is enlarged 1.2 times in the Y-axis direction. Therefore, the dust amount prediction unit 25 enlarges the region sandwiched between the side s4 and the opposite side ss4 1.2 times in the Y-axis direction to generate the Y component expansion part 324. Further, the side s2 is reduced to 0.66 times in the Y-axis direction. Therefore, the dust amount prediction unit 25 reduces the region sandwiched between the side s2 and the opposite side ss2 by 0.66 times in the Y-axis direction to generate the Y component reduction part 325.

As an example, when the dust amount prediction unit 25 expands the dust amount value in the Y-axis direction, a new element is added between two adjacent elements in the Y-axis direction, and the above-mentioned two adjacent elements are added thereto. Interpolate the mean value of. For example, the interpolated dust amount value may be a value obtained by rounding off the first decimal point of the average value.

As an example, when reducing the dust amount value in the Y-axis direction, the dust amount prediction unit 25 replaces two adjacent elements in the Y-axis direction with one element, and averages the above-mentioned two adjacent elements there. Assign a value. For example, the assigned garbage amount value may be a value obtained by rounding off the first decimal point of the average value.

(3) Generation of prediction map FIG. 11 is completed by the intermediate map 321 shown in FIG. 8 undergoing the above-mentioned steps of adjusting the amount of dust in the X-axis direction and the step of adjusting the amount of dust in the Y-axis direction. It is a figure which shows an example of the prediction map 33.

In the present embodiment, as an example, the dust amount prediction unit 25 has the Y component expansion part 324 and the Y component reduction part of the X component scaled map 331 obtained through the above-mentioned step of adjusting the dust amount in the X-axis direction. The prediction map 33 is generated in combination with 325.

In addition, when the Y component expansion part 324 and the Y component reduction part 325 are combined, the cells (coordinates) may overlap. In this case, for example, the dust amount prediction unit 25 may associate the average value of the dust amount value in the Y component expanding part 324 and the dust amount value in the Y component reducing part 325 of the overlapping squares with the square. ..

As described above, the dust amount prediction unit 25 completes the prediction map 33 by adjusting the distribution of the dust amount for each of the X component and the Y component of the intermediate map 321. In the above description, an example in which the amount of dust in the X-axis direction is first adjusted, and then the amount of dust in the Y-axis direction is adjusted based on the X component scaled map 331 obtained thereby. Is mentioned. However, the configuration of the dust amount prediction unit 25 is not limited to this. The dust amount prediction unit 25 first adjusts the dust amount in the Y-axis direction, generates an intermediate map in which the Y component has been scaled (not shown), and then, based on the intermediate map, the dust amount in the X-axis direction. May be adjusted.

The dust amount prediction unit 25 stores the prediction map 33 generated as described above in the storage unit 13. As a result, when the self-propelled vacuum cleaner 1 performs cleaning, even if the self-propelled vacuum cleaner 1 does not have information about the cleaning target area of the self-propelled vacuum cleaner 1, the cleaning control unit 21 can use the dust amount map having a similar floor surface shape. Cleaning can be performed efficiently with reference to the predicted prediction map 33.

(Processing flow)
FIG. 12 is a sequence diagram showing a flow of a process of predicting the amount of dust and cleaning, which is executed in the cleaning system.

In S101, the map acquisition unit 24 of the self-propelled vacuum cleaner 1 reads out the shape data 31 of its own room from the storage unit 13. For example, in response to the user instructing the start of cleaning via the operation panel 12, the map acquisition unit 24 executes the step of S101.

In S102, the map acquisition unit 24 generates search key data based on the shape data 31. The search key data is information indicating the feature amount of the floor surface shape indicated by the shape data 31, and has the same data structure as the processed shape data included in the processed dust amount map registered in the map database 51. Is generated.

In S103, the map acquisition unit 24 transmits the generated search key data to the cloud server 2 and requests the cloud server 2 for a reference map.

In S104, the map providing unit 41 of the cloud server 2 receives the search key data and accepts the request of the self-propelled vacuum cleaner 1.

In S105, the map providing unit 41 extracts the processed dust amount map having the received search key data and the processed shape data having the smallest error in the feature amount of the floor surface shape from the map database 51.

In S106, the map providing unit 41 transmits the extracted processed garbage amount map as a reference map to the self-propelled vacuum cleaner 1 of the requesting source.

In S107 (map acquisition step), the map acquisition unit 24 of the self-propelled vacuum cleaner 1 receives a reference map from the cloud server 2 via the communication unit 11, and stores the received reference map in the storage unit 13. The map acquisition unit 24 may restore the original dust amount map 34 based on the processed shape data and the dust amount value group data included in the reference map.

In S108, the dust amount prediction unit 25 rotates the reference map 32 stored in the storage unit 13 according to the shape data 31 of the own room, if necessary.

In S109, the dust amount prediction unit 25 deforms the floor surface shape shown in the reference map 32 according to the shape data 31. For example, the dust amount prediction unit 25 may enlarge or reduce the entire floor surface shape of the reference map 32 so as to match the floor surface shape of the shape data 31. Alternatively, for example, the dust amount prediction unit 25 may perform necessary enlargement or reduction for each side of the floor surface shape of the reference map 32 so as to match the floor surface shape of the shape data 31. Then, the dust amount prediction unit 25 generates the intermediate map 321 shown in FIG.

In S110 (dust amount prediction step), the dust amount prediction unit 25 predicts the dust amount in its own room with reference to the reference map 32. Specifically, the dust amount prediction unit 25 determines the dust amount for each coordinate of the shape data 31 of the own room based on the dust amount for each coordinate of the reference map 32. At this time, the dust amount prediction unit 25 determines the dust amount for each coordinate of the own room by adjusting the distribution of the dust amount value of the intermediate map 321 in consideration of the scaling ratio of the floor surface shape of the intermediate map 321. .. For example, the adjustment of the distribution of the dust amount value may be performed separately for each of the scaling ratio in the X component and the scaling ratio in the Y component carried out in S109.

In S111, the dust amount prediction unit 25 generates a prediction map 33 in which the dust amount is plotted for each coordinate of the shape data 31, and stores it in the storage unit 13.

In S112 (cleaning control step), the cleaning control unit 21 executes cleaning based on the prediction map 33. For example, the cleaning control unit 21 may increase the suction force of the suction unit 16 while traveling in an area with a large amount of dust (for example, the dust amount value is 4 or more). Further, the cleaning control unit 21 may control the traveling unit 14 to slow down the traveling speed while traveling in an area having a large amount of dust. Further, the cleaning control unit 21 may control the brush unit 15 to increase the rotation speed of the brush while traveling in an area with a large amount of dust.

According to the above method, the self-propelled vacuum cleaner 1 can predict the amount of dust in its own room and generate a prediction map with reference to a map of the amount of dust having a floor surface shape similar to that of its own room. Then, the self-propelled vacuum cleaner 1 can execute a cleaning operation according to the amount of dust based on the prediction map, and as a result, even if the learning of the amount of dust in the cleaning target area is insufficient, the self-propelled vacuum cleaner 1 can perform the cleaning operation. It is possible to predict the amount of dust in the area to be cleaned and carry out efficient cleaning.

[Embodiment 2]
Other embodiments of the present invention will be described below. For convenience of explanation, the same reference numerals are given to the members having the same functions as the members described in the above-described embodiment, and the description thereof will not be repeated.

Hereinafter, the configuration for constructing the map database 51 will be described.

(Software configuration of self-propelled vacuum cleaner 3)
FIG. 13 is a block diagram showing a configuration of a main part of another self-propelled vacuum cleaner 3. Like the self-propelled vacuum cleaner 1, the self-propelled vacuum cleaner 3 includes a communication unit 11, an operation panel 12, a storage unit 13, a traveling unit 14, a brush unit 15, a suction unit 16, a collision detection unit 17, and dust. A detection unit 18 is provided. Further, the self-propelled vacuum cleaner 3 includes a control unit 60 instead of the control unit 10 of the self-propelled vacuum cleaner 1. The control unit 60 includes a cleaning control unit 21, a sensor control unit 22, and a shape data generation unit 23, similarly to the control unit 10 of the self-propelled vacuum cleaner 1, and further, as an example, a map generation unit 61. , And a map processing unit 62.

The map generation unit 61 generates a dust amount map 34 based on the operation results of the cleaning control unit 21 and the sensor control unit 22. Specifically, the map generation unit 61 plots the dust amount value measured by the sensor control unit 22 for each coordinate of the shape data 31 generated by the shape data generation unit 23, and the dust amount as shown in FIG. Generate map 34.

The map processing unit 62 processes the dust amount map 34 to generate a processed dust amount map having a data structure suitable for searching. The map processing unit 62 transmits the generated processed garbage amount map to the cloud server 2. As described above, the processed dust amount map includes the processed shape data showing the feature amount of the floor surface shape shown by the shape data 31 and the dust representing the dust amount value for each coordinate in a text format (for example, CSV format). It is composed of including quantity value group data.

(Processing flow)
FIG. 14 is a sequence diagram showing the flow of the process of collecting the dust amount map executed in the cleaning system.

In S201, the cleaning control unit 21 starts cleaning based on an instruction from the user or the like.

In S202, when the self-propelled vacuum cleaner 3 determines that the learning about the cleaning target area of the self-propelled vacuum cleaner 3 is not sufficient, the cleaning control unit 21 needs to grasp the floor surface shape of the cleaning target area in the wall cleaning mode. To carry out. Specifically, the cleaning control unit 21 grasps the wall edge based on the detection result of the collision detection unit 17, and controls the traveling unit 14 so as to travel only on the wall edge. Then, the cleaning control unit 21 measures the length of the outer circumference of the floor surface shape of the own room, and transmits the measurement result to the shape data generation unit 23.

In S203, the shape data generation unit 23 generates shape data 31 indicating the floor surface shape based on the running record of the cleaning control unit 21 and the length of the outer periphery of the floor surface shape. Every time the collision detection unit 17 detects a collision, the cleaning control unit 21 repeats rotation, advancement, and the like of the drive wheels of the traveling unit 14, so that the angle (turning angle) of the corner of the cleaning target area is grasped. The shape data generation unit 23 generates shape data 31 based on the corner angle grasped based on the running record of the cleaning control unit 21 and the length of the outer circumference of the floor surface shape.

In S204, when the shape data 31 is generated, the cleaning control unit 21 then shifts to the full surface cleaning mode. Specifically, the cleaning control unit 21 controls the traveling unit 14 so as to pass through the entire floor surface including the wall surface so that the cleaning target area grasped by the shape data 31 can be cleaned without bias. In the full cleaning mode, the sensor control unit 22 operates the dust detection unit 18 to measure the amount of dust and the like sucked by the suction unit 16. The sensor control unit 22 transmits the measured value of the dust amount measured by the dust detection unit 18 to the map generation unit 61 in association with the coordinates of the shape data 31 generated in S203.

In S205 (map generation step), the map generation unit 61 generates a dust amount map 34 based on the above-mentioned measurement results. Specifically, the map generation unit 61 plots the associated dust amount for each coordinate in the shape data 31, and completes the dust amount map 34. As an example, the map generation unit 61 divides the measured value of the amount of dust detected by the dust detection unit 18 into five stages according to a predetermined threshold value, and the map generation unit 61 divides the actual measurement value into five stages of dust according to the amount of dust. Plot the quantity value for each coordinate. In this embodiment, as described above, the larger the number, the larger the amount of dust.

In S206, the map processing unit 62 determines whether or not the learning of the floor surface shape of the own room and the amount of dust has been sufficiently performed. How the map processing unit 62 determines the completion of learning is not particularly limited. For example, the map processing unit 62 may determine the completion of learning based on the fact that the number of times of cleaning has reached a predetermined number of times, or that the user has performed an input operation indicating learning completion. When the map processing unit 62 determines that the learning is completed, the process proceeds from YES in S206 to S207. When the map processing unit 62 determines that the learning is insufficient, the process returns from NO in S206 to S201. When returning to S201, the self-propelled vacuum cleaner 3 repeats each process from S202 to S205 every time cleaning is performed.

In S207, the map processing unit 62 reads out the dust amount map 34 generated by sufficient learning from the storage unit 13.

In S208 (map processing step), the map processing unit 62 generates processed shape data based on the shape data 31 included in the dust amount map 34. The processed shape data has the same data structure as the search key data 313 shown in FIG. 4, and shows the feature amount of the floor surface shape shown by the shape data 31. The map processing unit 62 defines the vertices of the shape data 31, defines the angle type of the internal angle of each vertex, and acquires the distance between the vertices, as when the map acquisition unit 24 generates the search key data.

In S209 (map processing step), the map processing unit 62 extracts the dust amount value for each coordinate included in the dust amount map 34, and generates, for example, text format dust amount value group data.

In S210 (map processing step), the map processing unit 62 generates a processed dust amount map in which the processed shape data is associated with the dust amount value group data, and transmits the processed dust amount map to the cloud server 2.

In S211 the map registration unit 42 of the cloud server 2 receives the processed dust amount map from the self-propelled vacuum cleaner 3 via the communication unit 11.

In S212, the map registration unit 42 of the cloud server 2 registers the processed dust amount map received from the self-propelled vacuum cleaner 3 in the map database 51.

[Modification example]
One self-propelled vacuum cleaner generates a function to predict the amount of dust in the room of the self-propelled vacuum cleaner 1 and a processed dust amount map suitable for searching that the other self-propelled vacuum cleaner 3 has. It may have both a function of uploading to the cloud server 2 and the function of uploading to the cloud server 2.

In this case, the control unit of the self-propelled vacuum cleaner includes a cleaning control unit 21, a sensor control unit 22, a shape data generation unit 23, a map acquisition unit 24, a dust amount prediction unit 25, a map generation unit 61, and a map processing unit. 62 is provided.

[Example of implementation by software]
Control block of self-propelled vacuum cleaner (in particular, cleaning control unit 21, sensor control unit 22, shape data generation unit 23, map acquisition unit 24, dust amount prediction unit 25, map generation unit 61, and map processing unit 62) May be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or may be realized by software.

In the latter case, the self-propelled vacuum cleaner is equipped with a computer that executes the instructions of a program that is software that realizes each function. The computer includes, for example, at least one processor (control device) and at least one computer-readable recording medium that stores the program. Then, in the computer, the processor reads the program from the recording medium and executes the program, thereby achieving the object of the present invention. As the processor, for example, a CPU (Central Processing Unit) can be used. As the recording medium, a "non-temporary tangible medium", for example, a ROM (Read Only Memory) or the like, a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. Further, a RAM (Random Access Memory) for expanding the program may be further provided. Further, the program may be supplied to the computer via any transmission medium (communication network, broadcast wave, etc.) capable of transmitting the program. It should be noted that one aspect of the present invention can also be realized in the form of a data signal embedded in a carrier wave, in which the program is embodied by electronic transmission.

〔summary〕
In the self-propelled vacuum cleaner 1 according to the first aspect of the present invention, a plurality of dust amount maps 34 including the shape data 31 indicating the floor surface shape and the dust amount associated with each coordinate of the floor surface are registered. A map acquisition unit 24 that acquires a dust amount map including shape data similar to the target floor surface shape of the cleaning target area to be cleaned by the own machine as a reference map 32 from the map database 51, and the target floor surface. Based on the shape and the acquired reference map, the dust amount prediction unit 25 that predicts the dust amount for each coordinate of the target floor surface shape, and the predicted dust amount (prediction map 33), the self. It is provided with a cleaning control unit 21 that controls the cleaning operation of the machine.

According to the above configuration, it is possible to accurately predict the amount of dust on the floor of the own room by referring to the dust amount map of the room having another similar floor shape. As a result, even if the learning of the amount of dust on the floor surface shape of the own room is insufficient, it is possible to efficiently clean the own room based on the predicted amount of dust.

In the self-propelled vacuum cleaner 1 according to the second aspect of the present invention, in the first aspect, the map acquisition unit collects shape data indicating the target floor surface shape from a plurality of other self-propelled vacuum cleaners 3. The floor surface shape indicated by the shape data included in the dust amount map is similar to the target floor surface shape when transmitted to the server (cloud server 2) having the map database in which the dust amount map is registered. If so, the garbage amount map extracted from the map database by the server may be received from the server as the reference map.

According to the above configuration, the self-propelled vacuum cleaner is a room dust with a similar floor shape from a map database that stores a large amount of the results learned by many other self-propelled vacuum cleaners. A quantity map can be obtained as a reference map. As a result, the self-propelled vacuum cleaner can accurately predict the amount of dust on the floor of the room by referring to the reference map even if the learning of the amount of dust on the floor of the room is insufficient. Is possible.

In the self-propelled vacuum cleaner 1 according to the third aspect of the present invention, in the second aspect, the map acquisition unit transmits shape data indicating a feature amount extracted from the target floor surface shape to the server, and the feature is described. The quantity may include the number of vertices of the target floor surface shape, information on the angle of each vertex, and the distance between each vertex.

According to the above configuration, the server can compare features with simple data structures to determine the similarity between the target floor shape and the reference floor shape, and the server searches the map database. The processing load when doing this can be significantly reduced.

The self-propelled vacuum cleaner 1 according to the fourth aspect of the present invention includes a storage unit 13 for storing shape data 31 indicating the target floor surface shape in any one of the first to third aspects, and the dust amount prediction unit is provided. Based on the amount of dust for each coordinate indicated by the reference map, the target floor surface shape (shape shown in FIG. 3) and the reference floor surface shape (shape shown in FIG. 7) shown by the reference map. The amount of dust for each coordinate of the cleaning target area may be predicted by taking into account the difference between the above.

In the self-propelled vacuum cleaner 1 according to the fifth aspect of the present invention, in any one of the first to the fourth aspects, the dust amount prediction unit has the reference floor surface shape shown by the reference map matching the target floor surface shape. As described above, the scaling factor for scaling all or part of the reference floor surface shape is determined, and the distribution area of the dust amount for each coordinate arranged on the reference map is scaled based on the scaling factor. , The amount of dust for each coordinate of the cleaning target area may be predicted.

In this way, by deforming the distribution area of the dust amount value plotted on the shape data in the same manner as the floor surface shape, it is possible to accurately predict the dust amount for each coordinate of the cleaning target area.

In the self-propelled vacuum cleaner 1 according to the sixth aspect of the present invention, when the dust amount prediction unit determines in the fifth aspect that the target floor surface shape and the reference floor surface shape are similar to each other, the dust amount prediction unit determines that the target floor surface shape and the reference floor surface shape are similar. The scaling factor for enlarging or reducing the reference floor shape is determined so that the reference floor shape matches the target floor shape, and the target floor shape and the reference floor shape are similar to each other. If it is determined that the reference floor shape is not, the scaling ratio for enlarging or reducing each side forming the reference floor shape is determined for each side so that the reference floor shape matches the target floor shape. You may.

In this way, by enlarging or reducing the distribution area of the dust amount value plotted on the shape data in the same way as in the case of the floor surface shape, the dust amount for each coordinate of the cleaning target area is accurately predicted. be able to.

The self-propelled vacuum cleaner 3 according to the seventh aspect of the present invention is a dust amount map including shape data showing the floor surface shape of the cleaning target area of the own machine and the dust amount associated with each of the coordinates of the floor surface. 61, and a map processing unit 62 that processes the generated dust amount map in order to store the generated dust amount map in a map database in which the dust amount map collected from a plurality of self-propelled vacuum cleaners is registered. The map processing unit extracts the number of vertices of the floor surface shape, information on the angle of each vertice, and the distance between each vertice as a feature amount, and obtains the correspondence relationship between the coordinates and the dust amount value. The dust amount map is processed by generating the dust amount value group data shown in a format whose data size is smaller than that of the dust amount map.

According to the above configuration, the processed garbage amount map having a small data size, including the feature amount having a simple data structure, is registered in the map database. This can save a lot of storage space for storing the map database. Moreover, in the search of the map database, it is possible to compare the feature quantities having a simple data structure and determine the similarity between the target floor surface shape and the reference floor surface shape. It is possible to significantly reduce the processing load when a device such as a server that searches the map database searches the map database.

In the server (cloud server 2) according to the eighth aspect of the present invention, a plurality of dust amount maps including shape data indicating the floor surface shape and the dust amount associated with each coordinate of the floor surface are registered. Amount of dust including shape data similar to the target floor shape of the target area to be cleaned by the self-propelled vacuum cleaner with respect to the map database 51 and the self-propelled vacuum cleaner requesting the dust amount map. A map providing unit 41 that provides a map as a reference map is provided, and the map providing unit is registered in the map database and a first feature amount (search key data 313 in FIG. 4) of the target floor surface shape. Compared with the second feature amount of the floor surface shape (processed shape data 511 to 515, etc. in FIG. 5) shown by the shape data of the dust amount map, the second feature amount having the smallest error from the first feature amount. A dust amount map including shape data indicating the above is provided to the self-propelled vacuum cleaner as the reference map, and the first feature amount and the second feature amount are the number of vertices of the floor surface shape and the angle of each vertice. Includes information about and the distance between each vertex.

This enables the server to compare features with simple data structures and determine the similarity between the target floor shape and the reference floor shape, and the processing when the server searches the map database. The load can be significantly reduced.

In the control method of the self-propelled vacuum cleaner according to the ninth aspect of the present invention, a plurality of dust amount maps including the shape data indicating the floor surface shape and the dust amount associated with each coordinate of the floor surface are registered. A map acquisition step of acquiring a dust amount map including shape data similar to the target floor surface shape of the cleaning target area to be cleaned by the own machine as a reference map from the map database, the target floor surface shape, and the target floor surface shape. A dust amount prediction step that predicts the dust amount for each coordinate of the target floor surface shape based on the acquired reference map, and a cleaning control that controls the cleaning operation of the own machine based on the predicted dust amount. Including steps. As a result, the same effect as that of the self-propelled vacuum cleaner 1 according to the first aspect is obtained.

The control method of the self-propelled vacuum cleaner according to the tenth aspect of the present invention is dust including shape data indicating the shape of the floor surface of the cleaning target area of the own machine and the amount of dust associated with each coordinate of the floor surface. A map generation step for generating a quantity map and a map processing step for processing the generated dust amount map to be stored in a map database in which the dust amount map collected from a plurality of self-propelled vacuum cleaners is registered. In the map processing step, the number of vertices of the floor surface shape, information on the angle of each vertice, and the distance between each vertices are extracted as feature quantities, and the correspondence between the coordinates and the dust amount value is based on the above. The dust amount map is processed by generating the dust amount value group data shown in a format whose data size is smaller than that of the dust amount map. As a result, the same effect as that of the self-propelled vacuum cleaner 3 according to the seventh aspect is obtained.

The self-propelled vacuum cleaner according to each aspect of the present invention may be realized by a computer, and in this case, the self-propelled vacuum cleaner is operated by operating the computer as each part (software element) provided in the self-propelled vacuum cleaner. A control program for a self-propelled vacuum cleaner that realizes a vacuum cleaner with a computer, and a computer-readable recording medium that records the control program are also included in the scope of the present invention.

The present invention is not limited to the above-described embodiments, and various modifications can be made within the scope of the claims, and the embodiments obtained by appropriately combining the technical means disclosed in the different embodiments. Is also included in the technical scope of the present invention. Further, by combining the technical means disclosed in each embodiment, new technical features can be formed.

1 Self-propelled vacuum cleaner, 2 Cloud server (server), 3 Self-propelled vacuum cleaner, 10 Control unit, 11 Communication unit, 12 Operation panel, 13 Storage unit, 14 Travel unit, 15 Brush unit, 16 Suction unit, 17 Collision detection unit, 18 dust detection unit, 21 travel control unit, 22 sensor control unit, 23 shape data generation unit, 24 map acquisition unit, 25 dust amount prediction unit, 31 shape data, 32 reference map, 33 prediction map, 34 dust Quantity map, 40 control unit, 41 map provider unit, 42 map registration unit, 51 map database, 60 control unit, 61 map generation unit, 62 map processing unit, 100 cleaning system

Claims (12)

From the map database in which a plurality of dust amount maps including shape data indicating the floor surface shape and the dust amount associated with each floor surface coordinate are registered, the cleaning target area to be cleaned by the own machine is used. A map acquisition unit that acquires a dust amount map containing shape data similar to the target floor shape as a reference map,
A dust amount prediction unit that predicts the dust amount for each coordinate of the target floor surface shape based on the target floor surface shape and the acquired reference map.
A self-propelled vacuum cleaner equipped with a cleaning control unit that controls the cleaning operation of the own machine based on the predicted amount of dust. The map acquisition unit
The shape data indicating the target floor surface shape is transmitted to a server having the map database in which the dust amount map collected from a plurality of other self-propelled vacuum cleaners is registered.
Assuming that the floor surface shape indicated by the shape data included in the dust amount map is similar to the target floor surface shape, the dust amount map extracted from the map database by the server is used as the reference map. The self-propelled vacuum cleaner according to claim 1, which is received from the server. The map acquisition unit transmits shape data indicating a feature amount extracted from the target floor surface shape to the server.
The self-propelled vacuum cleaner according to claim 2, wherein the feature amount includes the number of vertices of the target floor surface shape, information on the angle of each vertex, and the distance between each vertex. A storage unit for storing shape data indicating the shape of the target floor surface is provided.
The dust amount prediction unit is based on the dust amount for each coordinate indicated by the reference map, and adds the difference between the target floor surface shape indicated by the shape data and the reference floor surface shape indicated by the reference map. The self-propelled vacuum cleaner according to any one of claims 1 to 3, which predicts the amount of dust for each coordinate of the area to be cleaned. The dust amount prediction unit determines the scaling ratio for scaling all or part of the reference floor shape so that the reference floor shape indicated by the reference map matches the target floor shape.
Any of claims 1 to 4, wherein the dust amount distribution area for each coordinate arranged on the reference map is scaled based on the scaling ratio to predict the dust amount for each coordinate of the cleaning target area. The self-propelled vacuum cleaner described in item 1. The garbage amount prediction unit is
When it is determined that the target floor surface shape and the reference floor surface shape are similar, the reference floor surface shape is enlarged or reduced so that the reference floor surface shape matches the target floor surface shape. Determine the scaling factor for
When it is determined that the target floor surface shape and the reference floor surface shape are not similar, each side forming the reference floor surface shape is enlarged so that the reference floor surface shape matches the target floor surface shape. The self-propelled vacuum cleaner according to claim 5, wherein the scaling ratio for scaling is determined for each side. A map generator that generates a dust amount map including shape data indicating the floor shape of the area to be cleaned of the own machine and the dust amount associated with each floor coordinate.
It is provided with a map processing unit that processes the generated dust amount map to store it in a map database in which the dust amount map collected from a plurality of self-propelled vacuum cleaners is registered.
The map processing unit
Information on the number of vertices of the floor surface shape, the angle of each vertex, and the distance between each vertex are extracted as feature quantities.
Self-propelled cleaning that processes the dust amount map by generating dust amount value group data that shows the correspondence between the coordinates and the dust amount value in a format whose data size is smaller than the original dust amount map. Machine. A map database in which a plurality of dust amount maps including shape data indicating the floor surface shape and the dust amount associated with each floor coordinate are registered, and
For a self-propelled vacuum cleaner that requests the dust amount map, a dust amount map containing shape data similar to the target floor surface shape of the cleaning target area to be cleaned by the self-propelled vacuum cleaner is used as a reference map. Equipped with a map provider to provide
The map provider
The first feature amount of the target floor surface shape is compared with the second feature amount of the floor surface shape indicated by the shape data of the dust amount map registered in the map database.
A dust amount map including shape data showing the second feature amount having the smallest error from the first feature amount is provided to the self-propelled vacuum cleaner as the reference map.
The first feature amount and the second feature amount include information on the number of vertices of the floor surface shape, information on the angle of each vertex, and the distance between each vertex. From the map database in which a plurality of dust amount maps including shape data indicating the floor surface shape and the dust amount associated with each floor surface coordinate are registered, the cleaning target area to be cleaned by the own machine is used. A map acquisition step to acquire a dust amount map containing shape data similar to the target floor surface shape as a reference map, and
A dust amount prediction step for predicting the dust amount for each coordinate of the target floor surface shape based on the target floor surface shape and the acquired reference map.
A method of controlling a self-propelled vacuum cleaner, including a cleaning control step that controls the cleaning operation of the vacuum cleaner based on the predicted amount of dust. A map generation step to generate a dust amount map including shape data showing the floor surface shape of the cleaning target area of the own machine and the dust amount associated with each floor coordinate.
A map processing step for processing the generated dust amount map to be stored in a map database in which the dust amount map collected from a plurality of self-propelled vacuum cleaners is registered is included.
In the map processing step,
Information on the number of vertices of the floor surface shape, the angle of each vertex, and the distance between each vertex are extracted as feature quantities.
Self-propelled cleaning that processes the dust amount map by generating dust amount value group data that shows the correspondence between the coordinates and the dust amount value in a format whose data size is smaller than the original dust amount map. How to control the machine. The control program for operating a computer as a self-propelled vacuum cleaner according to claim 1, wherein the computer functions as the map acquisition unit, the dust amount prediction unit, and the cleaning control unit. The control program for operating a computer as a self-propelled vacuum cleaner according to claim 7, wherein the computer functions as the map generation unit and the map processing unit.

Images