ES2914891A1 - Cleaning and/or Disinfection Robot with textile recognition and method to operate it (Machine-translation by Google Translate, not legally binding) - Google Patents

Abstract

Cleaning and/or disinfection robot that includes, at least, a body, means of measurement of surface Because the robot makes a map of the functional work area through the surface measurement means that, in combination with the means of image recognition, allows to establish and identify textile objects. In addition, the invention presents a method to operate the cleaning and/or disinfection robot, so that it recognizes limiting or passing textile objects. (Machine-translation by Google Translate, not legally binding)

Description

CLEANING AND/OR DISINFECTION ROBOT WITH MEANS OF RECOGNITION

TEXTILE AND METHOD TO OPERATE IT

TECHNICAL SECTOR

The present invention falls within cleaning and/or disinfection apparatus and, more specifically, in artificial intelligence systems with textile recognition means.

BACKGROUND OF THE INVENTION

Currently, cleaning and/or disinfection robots are mainly formed by a suction motor that sucks up the dirt that it collects by means of a central brush and finally leads it inside a waste tank; These devices are usually equipped with autonomous displacement means and surface measurement means, in addition to an information processing operating unit, in charge of processing all the information that operates its artificial intelligence means; combining movement logics with artificial vision and deep neural training of this type of apparatus belonging to the state of the art;

In this sense, we find the document EP3508937, which comprises a mobile cleaning robot that includes a cleaning head configured to clean the surface of the floor in an environment and at least one camera with a certain angle that has a field of vision that is extends above the ground surface. The at least one camera is configured to capture images that include portions of the environment above the ground surface. The robot includes a recognition module that is configured to recognize objects in the environment based on the images captured by the at least one camera, in which the recognition module is trained at least in part using the images captured by the at least one camera. a camera. The robot includes a storage device that is configured to store a map of the environment. The robot includes a control module configured to control the mobile cleaning robot to navigate in the environment using the map and to operate the cleaning head to perform cleaning tasks taking into account objects recognized by the recognition module.

On the other hand, we find document EP3234839A1, which comprises an autonomous mobile robot that includes a drive configured to maneuver the robot on a ground surface within an operating environment; a camera mounted on the robot having a field of view that includes the ground adjacent to the mobile robot in the driving direction of the mobile robot; a frame buffer that stores the image frames obtained by the camera while the mobile robot is driving; and a memory device configured to store a set of data learned from a plurality of descriptors corresponding to pixel patches in image frames corresponding to portions of the operating environment and determined by mobile robot sensor events.

We have not found a method similar to ours or one that effectively solves the technical problem regarding the failure of artificial intelligences in cleaning and/or suction robots to effectively recognize the different types of fabrics, and act accordingly.

EXPLANATION OF THE INVENTION

Cleaning and/or disinfection devices with autonomous or semi-autonomous means of movement have an increasingly extensive sensor system as technology advances. That is why artificial intelligence is already part of these robots. It is therefore the integration of at least one camera located on the front of the robot and that adapts to its morphology with a reduced size and that, as it moves forward, recognizes the objects that it finds in its path. In addition to taking into account the position, we must consider some characteristics of image resolution and related to color so that the frames that are taken resemble reality as much as possible.

The cleaning and/or disinfection robot comprises an artificial intelligence module, information storage means, and at least one camera, which must be interconnected by the base plate that is included inside the robot. In these information storage media we will store the algorithms of each object to be identified and, through the frames taken in real time, we will generalize the knowledge to categorize the identified object. Also included on the board based on a control module configured to make a series of decisions that will endow the robot with artificial intelligence.

To increase the accuracy in image recognition, a database categorized by types of movable property with textile characteristics that prevent the robot from performing a total cleaning of the surface is required. These movable goods with textile characteristics can be curtains, sheets, bedspreads, quilts, bedspreads, covers and/or tablecloths, among others, which practically reach the ground and the robot, when using surface measurement means, can confuse them with elements limiting.

In general, the furniture mentioned above often prevents the robot from cleaning underneath it, since, when detected by the surface measurement means, the usual result in the state of the art is that the artificial intelligences of said cleaning and/or disinfection robots recognize and interpret these textile objects, in limitations similar to a wall, delimiting the work area of the cleaning and/or disinfection robot; when in reality you could pass through these objects, which mostly make up a piece of furniture with an apparent textile structure, which the robot recognizes as walls, but thanks to the present invention, the robots will recognize it as a textile object and will try to overcome the false limits created by the wrong recognition of a limiting element such as a wall.

To confirm that this detection is correct and to make sure that there is really no wall or limiting element on the other side, we will need the verification of a third mechanical and/or electronic element. This element can be, for example, the bumper or also called the bumper that forms part of the front part of the robot vacuum cleaner, which collects the blows and indicates that there is an obstacle and, therefore, it is not possible to proceed. Thus, with these characteristics, we will be able to follow the navigation logics programmed in the robot to continue cleaning normally without having to move or lift the false limitation elements.

In a possible mode of operation, the invention comprises the following steps:

- Activation of the cleaning and/or disinfection robot, with the consequent activation of the surface measurement means and the artificial vision, to detect and check the elements that delimit the functional work area.

- The robot accesses its database to identify whether it is a previously categorized object with a textile appearance, and if so, perform an action by the cleaning and/or disinfection robot.

- Prior deep neural training, with a database of at least twenty thousand images for each recognition category with different objects of the same category in different environments.

- Basic training of textile objects that do not constitute a limiting element, which includes at least curtains, bedspreads, sheets, covers, tablecloths, bedspreads, quilts and blankets.

- Having detected a limiting element, the robot checks if it is a textile object through the image recognition means and; if so, react accordingly.

- The robot, being located in front of the textile object, will advance towards it with the intention of overcoming the limiting barrier, thus verifying that it is, going through it to continue cleaning the surface that was covered by this false limiting element;

- Otherwise, the robot checks that the textile object offers resistance to its passage and forms part of the limiting elements of the functional work area through its means of detecting objects by physical contact, then the robot again categorizes said textile object as a limiting element of the effective work area and continues with the navigation until finding another similar case to repeat the operation logic.

BRIEF DESCRIPTION OF THE DRAWINGS

To complement the description that is being made and in order to help a better understanding of the characteristics of the invention, a set of drawings is attached as an integral part of said description, where for illustrative and non-limiting purposes, the following has been represented: Next:

Figure 1.- Shows an operating diagram of the cleaning and/or disinfection robot, as well as the navigation logics that operate the cleaning and/or disinfection robot.

PREFERRED EMBODIMENT OF THE INVENTION

In the following detailed description of the preferred embodiments, reference is made to the accompanying drawings which form a part of this specification, and in which there are shown, by way of illustration, specific preferred embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention, and it is understood that other embodiments may be used and logical structural, mechanical, electrical, and/or chemical changes may be made without departing from the scope of the invention. To avoid details not necessary to enable those skilled in the art to carry out the detailed description should therefore not be taken in a limiting sense.

In a preferred embodiment of the invention, we propose the most common scenarios for which a cleaning and/or disinfection robot with surface measurement means, commonly through laser technology, could not access, due to the problem of detecting limiting elements by surface measurement means, but thanks to the integration of artificial intelligence it is possible. For this preferred embodiment, the most common places that we can find inside the home are the space between the curtains and the wall; the space found under a table when the tablecloth practically reaches the floor; and the space found under the bed when the quilt, comforter, or the like is substantially flush with the floor. To detect these categories, we will have a previously acquired database to compare them with the images that are captured in real time. In this way we will be able to distinguish the scenarios and carry out a reaction of the robot accordingly. The cleaning process that this recognition entails consists of a third element that will verify that the captured image falls within the characteristics mentioned above. This third factor must be at least one mechanical element, such as a bumper, located on the front of the robot vacuum cleaner. The function of the bumper is to warn the robot that it cannot move in that direction and therefore the cleaning and/or disinfection robot continues with the cleaning and/or disinfection operation in progress.

In another preferred embodiment, the invention comprises the following operating method:

- Activation of the cleaning and/or disinfection robot, with the consequent activation of the surface measurement means and the artificial vision camera, to detect and check the elements that delimit the functional work area.

- The robot accesses its storage device to identify if it is a previously categorized object or piece of furniture with a textile appearance, and if so, perform an action given by the previously configured control module.

- Prior deep neural training, with a database of at least twenty thousand images for each recognition category with different objects of the same category in different environments.

- Basic training of textile objects that do not constitute a limiting element, which includes at least curtains, bedspreads, sheets, covers, tablecloths, bedspreads, quilts and blankets.

- Having detected a limiting element, the robot checks if it is a textile object through the image recognition means and; if so, react accordingly.

- The robot, being located in front of the object categorized as textile or textile furniture, will move towards it with the intention of overcoming the limiting barrier, thus verifying that it is.

- If the robot can overcome the textile object, it discards it as a limiting element and, consequently, it crosses it and continues cleaning the surface that was covered by this false limiting element, adding the new area to the existing map;

- Otherwise, the robot checks that the textile object offers resistance to its passage and forms part of the limiting elements of the functional work area, through its means of detecting objects by physical contact, then the robot categorizes again said textile object as a limiting element of the effective work area and continues with the navigation until finding another similar case to repeat the logic.

Claims (3)

1. Cleaning and/or disinfection robot that includes, at least, - a body; - surface measuring means; - means of artificial vision; - image recognition means; - autonomous and/or semi-autonomous means of movement; Y, - physical obstacle detection means; Characterized because, - The robot makes a map of the functional work area through surface measurement means that, in combination with image recognition means, allows establishing and identifying textile objects; - The robot makes use of its physical obstacle detection means to check if each of the textile objects identified on the previously configured map are limiting elements; - Its information operative unit is configured to create a map of the work area, identify the textile objects and check the condition of the textile object and corroborate the possibility of passing the cleaning and/or disinfection robot through the textile object, accessing the surface previously covered and adding the new area identified in the current map of the effective work surface. 2. Method for operating a cleaning and/or disinfection robot with textile recognition means comprising, at least, the following steps; - the cleaning and/or disinfection robot activates its surface measurement means, which perform the function of mapping the effective work area and detect the limiting elements, creating a map of the work surface; - the cleaning and/or disinfection robot activates its image recognition means, recognizing and locating limiting elements, from a predefined visual database; - the cleaning and/or disinfection robot checks the limiting elements, through image recognition means, while the robot moves through the effective work area in a cleaning operation, identifying textile objects, among the limiting elements detected; - The cleaning and/or disinfection robot checks the textile objects through the means of physical detection of obstacles, to find out if they are textile objects with passage through or not, with the following options: - if the textile object does not allow passage through it because it is not a pass-through textile element, the robot categorizes it as a limiting element and represents the new limiting area identified in the current map of the effective work surface ; - if the textile object allows passage through it, because it is a pass-through textile element, the robot crosses it and accesses the surface previously delimited by the at least limiting elements; and now categorized as a pass-through textile element, a passage through which the robot accesses an area not recognized by its surface measurement means, recognizes the work area and adds it to the map of the rest of the room. 3. Textile recognition method through image recognition means for a cleaning and/or disinfection robot comprising at least the following steps; - Activation of the cleaning and/or disinfection robot, with the consequent activation of the surface measurement means and the artificial vision camera, to detect and check the elements that delimit the functional work area. - The robot accesses its storage device to identify if it is a previously categorized object or piece of furniture with a textile appearance, and if so, perform an action given by the previously configured control module. - Prior deep neural training, with a database of at least twenty thousand images for each recognition category with different objects of the same category in different environments. - Basic training of textile objects that do not constitute a limiting element, which includes at least curtains, bedspreads, sheets, covers, tablecloths, bedspreads, quilts and blankets. - Having detected a limiting element, the robot checks whether it is a textile object through image recognition means Y; if so, react accordingly. - The robot, being located in front of the object categorized as textile or textile furniture, will move towards it with the intention of overcoming the limiting barrier, thus verifying that it is. - If the robot can overcome the textile object, it discards it as a limiting element and, consequently, it crosses it and continues cleaning the surface that was covered by this false limiting element, adding the new area to the existing map. - Otherwise, the robot checks that the textile object offers resistance to its passage and forms part of the limiting elements of the functional work area, through its means of detecting objects by physical contact, then the robot categorizes again said textile object as a limiting element of the effective work area and continues with the navigation until finding another similar case to repeat the logic. 1