Method for generating indoor map, and device for performing same

Abstract

The present invention relates to a map generation method performed by a processor of an indoor map generation device. The map generation method comprises the steps of: obtaining a point cloud corresponding to an indoor space including a static obstacle and a dynamic obstacle; clustering the point cloud to generate a map for the indoor space including the static obstacle; and obtaining the point cloud corresponding to the indoor space a plurality of times to update points matched to the map.

Description

Method for generating an indoor map and a device for performing the same

[0001] The present invention relates to a method for generating an indoor map and an apparatus for performing the same. More specifically, it relates to a method for accurately generating a map of the interior of a building including a glass structure and an apparatus for performing the same.

[0002] With the advancement of Autonomous Mobile Robot (AMR) technology, interest in and the need for AMRs are increasing day by day across all industries. As AMRs perceive their surroundings using LiDAR sensors and can autonomously plan movement paths and perform tasks based on this, their use in commercial and industrial buildings is becoming widespread for purposes such as building management, logistics, and work automation.

[0003] As such, the Simultaneous Localization and Mapping (SLAM) algorithm, which is one of the core technologies of autonomous robots, is a technology that estimates the robot's own position while simultaneously generating a map of the robot's surrounding environment. By reflecting changes in the surrounding environment in real time, it enables the robot to perform tasks stably even in environments with obstacles.

[0004] However, unlike opaque objects, transparent objects that exhibit strong specular reflection, such as glass windows inside buildings, allow lasers emitted from LiDAR sensors to pass through and possess light reflection characteristics that cause noise, posing an obstacle to the mission performance of autonomous robots. As a result, autonomous robots perceive areas where glass walls are located as empty spaces and move accordingly, leading to problems where they collide with the glass walls.

[0005] The background description of the invention is provided to facilitate a better understanding of the present invention. The matters described in the background description should not be construed as an acknowledgment that they exist as prior art.

[0006] Accordingly, beyond the conventional method of equipping autonomous robots with only 2D LiDAR sensors, there was an attempt to recognize transparent objects such as glass by equipping them with ultrasonic sensors, cameras, etc. However, since fusing various sensor data acquired from autonomous robots requires new hardware and software, development costs increase compared to using a single 2D LiDAR sensor.

[0007] Meanwhile, even when using a single 2D LiDAR sensor, there are limitations to commercialization because it utilizes high-end 2D LiDAR sensors that have high resolution, a wide scan range, and a high number of scans per second. In particular, because existing low-end 2D LiDAR sensors have low resolution, a narrow scan range, and a low number of scans per second, it is not possible to utilize the transparent object identification technique using intensity, which is used in high-end 2D LiDAR sensors.

[0008] Accordingly, a new method is required to generate maps of buildings containing glass structures more accurately using an autonomous robot equipped with a low-cost 2D LiDAR sensor.

[0009] As a result, the inventors of the present invention sought to develop a method to obtain a point cloud defined as the surface of an obstacle by adjusting the parameter settings of points collected from an autonomous driving robot equipped with a low-cost 2D LiDAR sensor, and to distinguish obstacles having transparent attributes through clustering.

[0010] In particular, the inventors of the present invention have configured a method to clearly distinguish between dynamic (or temporary) obstacles and static obstacles by accumulating and collecting point clouds and repeatedly updating them.

[0011] The problems of the present invention are not limited to those mentioned above, and other unmentioned problems will be clearly understood by those skilled in the art from the description below.

[0012] To solve the problem described above, an indoor map generation method according to an embodiment of the present invention is provided. The method is a map generation method performed by a processor of an indoor map generation device, and is configured to include the steps of: acquiring a point cloud corresponding to an indoor space containing static obstacles and dynamic obstacles; clustering the point cloud to generate an occupancy grid map for the indoor space containing the static obstacles; and acquiring the point cloud corresponding to the indoor space multiple times and updating the points matched to the occupancy grid map.

[0013] According to a feature of the present invention, the step of acquiring the point cloud may be a step of acquiring a point cloud corresponding to the indoor space, wherein the point cloud excluding the free space within the indoor space is acquired.

[0014] According to a feature of the present invention, the step of acquiring the point cloud may further include the step of identifying the type of sensor included in the device and the step of adjusting the map generation parameters of the device according to the type of sensor.

[0015] According to a feature of the present invention, the step of acquiring the point cloud may be a step of acquiring a point cloud corresponding to an indoor space that includes an obstacle having a transparent attribute among the static obstacles based on the adjustment of the map generation parameters.

[0016] According to a feature of the present invention, the step of generating the occupancy grid map may further include the step of extracting discontinuous regions from the point cloud and the step of clustering the discontinuous regions to remove clusters smaller than a preset size.

[0017] According to a feature of the present invention, the step of updating the points may further include the step of acquiring a point cloud a predetermined number of times at a predetermined time interval and accumulating points in the point cloud, and the step of distinguishing the static obstacle and the dynamic obstacle in the accumulated point cloud.

[0018] According to a feature of the present invention, the step of distinguishing between the static obstacle and the dynamic obstacle may further include the step of calculating the probability that a static obstacle exists at each point constituting the accumulated point cloud.

[0019] According to a feature of the present invention, the step of calculating the probability may further include the step of determining the weight of each of the N point clouds obtained a preset number of times and the step of determining whether the sum of the weights of each of the N point clouds is greater than or equal to a preset value.

[0020] To solve the problem described above, an indoor map generating device according to another embodiment of the present invention is provided. The device comprises a communication interface, a memory, and a processor operably connected to the communication interface and the memory. The processor is configured to acquire a point cloud corresponding to an indoor space containing static obstacles and dynamic obstacles, cluster the point cloud to generate an occupancy grid map for the indoor space containing static obstacles, acquire the point cloud corresponding to the indoor space multiple times, and update the points matched to the occupancy grid map.

[0021] Specific details of other embodiments are included in the detailed description and drawings.

[0022] The present invention enables an autonomous driving robot comprising only a single 2D LiDAR sensor to recognize transparent obstacles such as glass windows and glass walls. The present invention can generate an indoor map that recognizes transparent obstacles without using inefficient methods such as manually labeling spaces where transparent obstacles are placed after map generation or restricting entry itself, while also maintaining the performance of the SLAM algorithm.

[0023] The present invention does not require the installation of expensive 2D LiDAR sensors, fusion sensors, or complex neural network algorithms using fusion sensors, thereby reducing the development costs required to generate building maps using autonomous robots.

[0024] The present invention can clearly distinguish between dynamic and static obstacles within a building through the cumulative collection of point clouds acquired by an autonomous robot, rather than utilizing specific patterns occurring in transparent obstacles. For example, even with an autonomous robot equipped only with low-cost 2D LiDAR sensors, it is possible to generate an accurate indoor map even in spaces with many temporary static or dynamic obstacles, such as department stores, airports, and large supermarkets.

[0025] The effects according to the present invention are not limited to those exemplified above, and a wider variety of effects are included within the present invention.

[0026] FIG. 1 is a schematic diagram illustrating an indoor map generation method according to one embodiment of the present invention.

[0027] FIG. 2 is a block diagram showing an indoor map generation system according to one embodiment of the present invention.

[0028] FIG. 3 is a block diagram showing the configuration of an indoor map generating device according to one embodiment of the present invention.

[0029] FIG. 4 is a schematic flowchart of an indoor map generation method according to one embodiment of the present invention.

[0030] FIG. 5 is an exemplary diagram illustrating a point cloud collected by an indoor map generating device according to one embodiment of the present invention.

[0031] FIG. 6 is a detailed flowchart of an indoor map generation method according to one embodiment of the present invention.

[0032] FIG. 7 is an exemplary diagram illustrating a map produced by an indoor map generating device according to one embodiment of the present invention.

[0033] The advantages and features of the present invention and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below but may be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. In connection with the description of the drawings, similar reference numerals may be used for similar components.

[0034] In this document, expressions such as "have," "can have," "include," or "can include" refer to the existence of the relevant feature (e.g., numerical values, functions, actions, or components, etc.) and do not exclude the existence of additional features.

[0035] In this document, expressions such as “A or B,” “at least one of A or / and B,” or “one or more of A or / and B” may include all possible combinations of items listed together. For example, “A or B,” “at least one of A and B,” or “at least one of A or B” may refer to cases including (1) at least one A, (2) at least one B, or (3) both at least one A and at least one B.

[0036] Expressions such as "first," "second," "first," or "second" used in this document may modify various components regardless of order and / or importance, and are used merely to distinguish one component from another without limiting such components. For example, the first user device and the second user device may represent different user devices regardless of order or importance. For example, without departing from the scope of rights set forth in this document, the first component may be named the second component, and similarly, the second component may be renamed the first component.

[0037] Where it is stated that a certain component (e.g., a first component) is "(operatively or communicatively) coupled with" or "connected to" another component (e.g., a second component), it should be understood that the said certain component may be directly connected to the said other component or connected through another component (e.g., a third component). On the other hand, where it is stated that a certain component (e.g., a first component) is "directly connected" or "directly connected" to another component (e.g., a second component), it may be understood that no other component (e.g., a third component) exists between the said certain component and the said other component.

[0038] As used in this document, the expression “configured to” may be replaced, depending on the context, with, for example, “suitable for,” “having the capacity to,” “designed to,” “adapted to,” “made to,” or “capable of.” The term “configured to” does not necessarily mean only that which is “specifically designed to” in hardware. Instead, in some situations, the expression “device configured to” may mean that the device is “capable of” in conjunction with other devices or components. For example, the phrase “processor configured to perform A, B, and C” may mean a dedicated processor for performing those operations (e.g., an embedded processor), or a generic-purpose processor (e.g., a CPU or application processor) capable of performing those operations by executing one or more software programs stored in a memory device.

[0039] The terms used in this document are used merely to describe specific embodiments and are not intended to limit the scope of other embodiments. Singular expressions may include plural expressions unless the context clearly indicates otherwise. Terms used herein, including technical or scientific terms, may have the same meaning as generally understood by those skilled in the art described in this document. Terms used in this document that are defined in general dictionaries may be interpreted as having the same or similar meaning as they have in the context of the relevant art, and are not to be interpreted in an ideal or overly formal sense unless explicitly defined in this document. In some cases, even terms defined in this document may not be interpreted to exclude the embodiments of this document.

[0040] The features of each of the various embodiments of the present invention may be combined or combined with one another, either partially or wholly, and as will be fully understood by those skilled in the art, various technical interlocking and operation are possible, and each embodiment may be implemented independently of one another or together in an interlocking relationship.

[0041] Hereinafter, the present invention will be described in detail by explaining preferred embodiments of the present invention with reference to the attached drawings.

[0042] FIG. 1 is a schematic diagram illustrating an indoor map generation method according to one embodiment of the present invention.

[0043] Referring to FIG. 1, an indoor map generation method according to an embodiment of the present invention can generate an indoor map generation device (100) while driving through an indoor space, that is, a movable area within the indoor space, i.e., an indoor map. The indoor map generation device (100) can generate a map while simultaneously estimating its position in an indoor space containing an obstacle (11). The indoor map generation device (100) can recognize obstacles (11) using a LiDAR sensor, but the LiDAR sensor here has the disadvantage that while it can detect opaque obstacles, it is difficult to detect transparent obstacles. Accordingly, an indoor map generation device (100) equipped only with a LiDAR sensor can recognize obstacles (12) such as opaque doors, but has limitations in recognizing obstacles (13) such as transparent glass walls. In particular, low-cost LiDAR sensors, which have difficulty utilizing laser intensity, have difficulty obtaining detectable patterns through transparent obstacles.

[0044] An indoor map generation method according to one embodiment of the present invention can generate an indoor movement path using only data recognized by a low-cost LiDAR sensor, which has difficulty utilizing laser intensity. Here, a low-cost LiDAR sensor refers to a two-dimensional measurement sensor having a minimum angular resolution of 1.5 degrees or a maximum scan frequency of 10 Hz. In the present invention, even when using a single low-spec LiDAR sensor having low angular resolution and a small scan frequency value, it is possible to generate an indoor movement path that recognizes static obstacles, dynamic obstacles, and even transparent obstacles.

[0045] FIG. 2 is a block diagram showing an indoor map generation system according to one embodiment of the present invention.

[0046] Referring to FIG. 2, an indoor map generation system (1000) may include an indoor map generation device (100) that moves through an indoor space and generates a map of the indoor space, and a user device (200) that manages the indoor map generation device (100).

[0047] The indoor map generation device (100) may be a device capable of determining its own location, recognizing obstacles, and moving within an indoor space, or it may be a device of a service server that provides a service to enable such a device to generate a map of an indoor space. Specifically, the indoor map generation device (100) may perform map creation and location estimation simultaneously using a Simultaneous Localization and Mapping (SLAM) algorithm. For example, the indoor map generation device (100) may be a commercial robot such as a cleaning robot or a service robot, or an industrial robot such as a logistics robot. As another example, the indoor map generation device (100) may include a general-purpose computer, a laptop, and a data server capable of providing indoor space map generation services.

[0048] In various embodiments, the indoor map generating device (100) may obtain initial data for generating an indoor movement path from a user device (200). Here, the initial data may include an initial location where a robot is located within an indoor space. The indoor map generating device (100) may primarily generate a map based on the initial location using a SLAM algorithm, and subsequently iteratively update the map according to the indoor map generating method of the present invention. The indoor map generating device (100) may provide the iteratively updated map data to the user device (200). Here, the map data may include a two-dimensional map in which the indoor map generating device (100) can move, excluding obstacles within the indoor space. Additionally, the iteratively updated map data may include a map in which static obstacles and dynamic (or temporary) obstacles existing in the indoor space are distinguished.

[0049] In various embodiments, the indoor map generating device (100) may acquire a point cloud for an indoor space containing static obstacles and dynamic obstacles. Specifically, the point cloud is data collected using a low-cost, low-spec LiDAR sensor included in the indoor map generating device (100), and may represent a set of two-dimensional points (or coordinate values) corresponding to the indoor space. Additionally, static obstacles may have attributes that are fixed at a specific location, while dynamic obstacles may have attributes that are not fixed at a specific location. Furthermore, static obstacles and dynamic obstacles may be classified based on the LiDAR sensor into opaque obstacles that reflect lasers and transparent obstacles that transmit lasers. For example, an obstacle with transparent attributes may be a glass door, and a dynamic obstacle may be a person or a person's footprint.

[0050] In various embodiments, each coordinate of the point cloud acquired by the indoor map generating device (100) corresponds to an indoor space, and free space within the indoor space may be excluded. Typically, when a laser emitted from a LiDAR sensor passes through, a conventional map generating device defines this as free space. However, since transparent obstacles such as glass doors also allow the laser to pass through, a device containing only a low-cost, low-spec LiDAR sensor may make an error by defining the location where the transparent obstacle is placed as free space. Accordingly, the indoor map generating device (100) of the present invention can recognize only obstacles excluding free space, without recognizing both obstacles and free space using a LiDAR sensor. For example, the indoor map generating device (100) may change the setting value for recognizing free space among the map generating parameters set in the SLAM algorithm so that the LiDAR sensor of the indoor map generating device (100) collects only the laser reflected from obstacles.

[0051] In various embodiments, the indoor map generating device (100) can identify the type of sensor included in the device. Specifically, the indoor map generating device (100) can identify whether the sensor included in the device is a single LiDAR sensor, and the angular resolution and scan frequency value of the LiDAR sensor. Accordingly, if the device includes only a low-spec LiDAR sensor, the indoor map generating device (100) can adjust the map generating parameters of the device. For example, the indoor map generating device (100) can change a parameter set so that any one point can be recognized as free space from 'true' to 'false'.

[0052] In this way, the indoor map generating device (100) can obtain a point cloud corresponding to an indoor space including a transparent obstacle by adjusting the map generating parameters. The point cloud here may include points obtained near the vertical angle of incidence of the glass surface. In other words, the indoor map generating device (100) can obtain points to distinguish the transparent obstacle as an obstacle without recognizing it as free space.

[0053] Meanwhile, the indoor map generating device (100) can generate an occupancy grid map using the acquired point cloud. In the occupancy grid map generated by the indoor map generating device (100) through a SLAM algorithm, each grid may contain the probability that an obstacle exists.

[0054] In various embodiments, the indoor map generating device (100) can cluster a point cloud to generate an occupancy grid map for an indoor space containing static obstacles. Specifically, as the indoor map generating device (100) collects all points through a LiDAR sensor, the point cloud may contain not only transparent obstacles but also dynamic (or temporary) obstacles. Accordingly, the indoor map generating device (100) can extract discontinuous areas from the point cloud to remove dynamic obstacles. In other words, the indoor map generating device (100) can extract areas where discontinuous points exist. The indoor map generating device (100) can cluster discontinuous areas and remove clusters smaller than a preset size. For example, the indoor map generating device (100) may remove discontinuous areas extracted by people, while not removing discontinuous areas extracted by glass doors, according to a preset size.

[0055] In various embodiments, the indoor map generating device (100) can acquire a point cloud for an indoor space multiple times to resolve obstacle recognition errors through clustering and update points matched to the map. Specifically, the indoor map generating device (100) can acquire a point cloud a predetermined number of times at a predetermined time interval and accumulate points in the point cloud. The indoor map generating device (100) can autonomously move within the indoor space based on the device's start position, end position, and direction data provided in advance by the user, thereby eliminating unnecessary tasks that require an administrator to monitor the indoor map generating device (100) in real time. The indoor map generating device (100) can acquire, for example, N point clouds (N is a natural number greater than or equal to 2) through autonomous movement, and can acquire the first to Nth point clouds through point accumulation. The indoor map generating device (100) can distinguish between static obstacles and dynamic obstacles in the accumulated point clouds.

[0056] In various embodiments, the indoor map generating device (100) can calculate the probability that a static obstacle exists at each point constituting the accumulated point cloud. For example, the indoor map generating device (100) can calculate the probability that a static obstacle exists for each grid in a grid map based on each of the first to Nth point clouds. The probability of a static obstacle existing is determined by the sum of the probability values ​​in each point cloud, and the indoor map generating device (100) can determine the weight for each of the N point clouds (i.e., the first to Nth point clouds) obtained a preset number of times. For example, the indoor map generating device (100) can apply a weight of 1 / (1+M) to each of the N point clouds according to the number of times the point clouds are obtained (M, where M is a natural number greater than or equal to 2). The indoor map generating device (100) can multiply the probability values ​​existing for each grid in each of the N point clouds by the weight, and can sum the multiplication results calculated for each grid. The indoor map generating device (100) can determine whether this sum value is a preset value. If the sum value is greater than or equal to the preset value, the indoor map generating device (100) can determine that there is a static obstacle in the grid.

[0057] Meanwhile, the value used as a criterion for determining the presence of static obstacles may vary depending on the environment. For example, since the probability of the presence of dynamic and static obstacles differs between environments with few dynamic obstacles (e.g., unmanned factories, automated warehouses, logistics centers, etc.) and environments with many dynamic obstacles (e.g., shopping malls, airports, etc.), the indoor map generating device (100) may receive a reference value from a user or adjust the reference value according to the ratio of continuous areas and discontinuous areas within the point cloud.

[0058] For example, unmanned factories, automated warehouses, and logistics centers may have few active people and may contain obstacles such as stationary machines or shelves. Additionally, even if there are static obstacles such as Automated Guided Vehicles (AGVs), they move along periodic and fixed paths. Accordingly, the indoor map generating device (100) can select a path without initial obstacle interference, avoid obstacles in real time, and move along a simple and fixed path. That is, the indoor map generating device (100) can generate an indoor map with a stable path, although the searchable indoor space is limited, by lowering the threshold value for determining the presence of obstacles in an environment with few dynamic obstacles and reducing the real-time obstacle avoidance sensitivity.

[0059] Conversely, for example, shopping malls and airports have a large number of active people, and dynamic obstacles may occur frequently. Accordingly, only obstacles that are detected identically by the indoor map generating device (100) in every number of repeated attempts may be considered as static obstacles, while other obstacles may be considered as dynamic obstacles and removed from the map. That is, the indoor map generating device (100) can generate an indoor map while adapting to dynamic changes and expanding the searchable indoor space by increasing the threshold value for determining the presence of obstacles in an environment with many dynamic obstacles and increasing the real-time obstacle avoidance sensitivity.

[0060] In various embodiments, the indoor map generating device (100) may receive a reference value from a user or evaluate the environment of dynamic obstacles in the indoor space (e.g., evaluating the environment based on the ratio of continuous and discontinuous areas within a point cloud) and adjust the reference value. That is, the indoor map generating device (100) can generate a map using a reference value suitable for the environment in each of an environment with many dynamic obstacles and an environment with few dynamic obstacles.

[0061] The user device (200) is a device possessed by a user who wishes to obtain an indoor map, and may include a smartphone, tablet PC (Personal Computer), laptop, PC, etc. The user device (200) may provide the device's initial location, end location, and direction data to the indoor map generating device (100). Additionally, the user device (200) may receive a map that is repeatedly updated from the indoor map generating device (100).

[0062] In various embodiments, when the indoor map generating device (100) is a service server that provides a service to generate an occupancy grid map of an indoor space, the indoor map generating device (100) may provide a web or mobile application capable of generating an indoor map to an autonomous driving device (not shown). The autonomous driving device may install or run the web or mobile application or program, and may generate an indoor map by receiving the device's start position, end position, and direction data from the user device (200).

[0063] Up to now, an indoor map generation system (1000) according to one embodiment of the present invention has been described. According to the present invention, the indoor map generation system (1000) can accurately distinguish static obstacles having transparent attributes and also distinguish between static obstacles and dynamic obstacles by accumulating and collecting point clouds corresponding to an indoor space using only a device having a low-spec and low-cost LiDAR sensor.

[0064] FIG. 3 is a block diagram showing the configuration of an indoor map generating device according to one embodiment of the present invention.

[0065] Referring to FIG. 3, the indoor map generating device (100) may include a communication interface (110), a memory (120), an I / O interface (130), a 2D LiDAR sensor (140), and a processor (150), and each component may communicate with one or more communication buses or signal lines.

[0066] The communication interface (110) can be connected to the user device (200) via a wired / wireless communication network to exchange data. For example, the communication interface (110) can receive initial data from the user device (200) including the initial location where the indoor map generating device (100) is located within the indoor space. Additionally, the communication interface (110) can transmit repeatedly updated map data to the user device (200).

[0067] Meanwhile, a communication interface (110) that enables the transmission and reception of such data includes a wired communication port (111) and a wireless circuit (112), wherein the wired communication port (111) may include one or more wired interfaces, for example, Ethernet, Universal Serial Bus (USB), FireWire, etc. Additionally, the wireless circuit (112) may transmit and receive data with an external device through an RF signal or an optical signal. Furthermore, wireless communication may use at least one of a plurality of communication standards, protocols, and technologies, such as GSM, EDGE, CDMA, TDMA, Bluetooth, Wi-Fi, VoIP, Wi-MAX, or any other suitable communication protocol.

[0068] The memory (120) can store various data used in the indoor map generation device (100). For example, the memory (120) can store a point cloud of the space currently located and a grid map generated based thereon. Additionally, the memory (120) can store initial data related to the initial location of the device. Additionally, the memory (120) can store reference values ​​for identifying obstacles.

[0069] In various embodiments, the memory (120) may include a volatile or non-volatile recording medium capable of storing various data, commands, and information. For example, the memory (120) may include at least one type of storage medium among a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), RAM, SRAM, ROM, EEPROM, PROM, network storage, cloud, and blockchain database.

[0070] In various embodiments, the memory (120) may store at least one configuration of an operating system (121), a communication module (122), a user interface module (123), one or more applications (124) or a sensor processing module (125).

[0071] An operating system (121) (e.g., embedded operating systems such as LINUX, UNIX, MAC OS, WINDOWS, VxWorks, etc.) may include various software components and drivers for controlling and managing general system operations (e.g., memory management, storage device control, power management, etc.) and may support communication between various hardware, firmware, and software components.

[0072] The communication module (122) can support communication with another device through the communication interface (110). The communication module (120) may include various software components for processing data received by the wired communication port (111) or wireless circuit (112) of the communication interface (110).

[0073] The user interface module (123) can receive requests or inputs from a user, such as a keyboard, touch screen, keyboard, mouse, microphone, etc., through the I / O interface (130) and provide a user interface on the display.

[0074] The application (124) may include a program or module configured to be executed by one or more processors (150). Here, the application for generating an indoor map may be implemented on a server farm.

[0075] The sensor processing module (125) can process sensor-related functions (e.g., processing reflection data obtained through the 2D LiDAR sensor (140)). For example, the sensor processing module (125) can measure the distance between the device and the obstacle based on the reflection data obtained through the 2D LiDAR sensor (140).

[0076] The I / O interface (130) can connect at least one of the input / output devices (not shown) of the indoor map generating device (100), such as a display, keyboard, touch screen, and microphone, to the user interface module (123). The I / O interface (130) can receive user input (e.g., voice input, keyboard input, touch input, etc.) together with the user interface module (123) and process commands based on the received input.

[0077] The 2D LiDAR sensor (140) can emit a laser to measure the distance to surrounding objects and receive signals reflected from surrounding objects. The 2D LiDAR sensor (140) can scan a two-dimensional plane by rotating the laser.

[0078] The processor (150) is connected to a communication interface (110), a memory (120), an I / O interface (130), and a 2D LiDAR sensor (140) to control the overall operation of the indoor map generating device (100), and can execute various commands to generate an accurate indoor map based on the 2D LiDAR sensor (140) through an application or program stored in the memory (120).

[0079] The processor (150) may correspond to a computing device such as a CPU (Central Processing Unit) or an AP (Application Processor). Additionally, the processor (150) may be implemented in the form of an Integrated Chip (IC), such as a System on Chip (SoC) that integrates various computing devices. Alternatively, the processor (150) may include a module for computing artificial neural network models, such as a Neural Processing Unit (NPU).

[0080] Hereinafter, with reference to FIG. 4, a method for the processor (150) of the indoor map generating device (100) to generate an indoor map will be described.

[0081] FIG. 4 is a schematic flowchart of an indoor map generation method according to one embodiment of the present invention.

[0082] Referring to FIG. 4, the processor (150) can acquire a point cloud for an indoor space containing static obstacles and dynamic obstacles (S110). Specifically, the point cloud is data collected using a low-cost, low-spec 2D LiDAR sensor (140) included in the indoor map generating device (100), and may represent a set of 2D points (or coordinate values) corresponding to the indoor space. The processor (150) can accumulate points collected in the indoor space while driving in the indoor space.

[0083] Meanwhile, static obstacles have properties that are fixed at a specific location, whereas dynamic obstacles may have properties that are not fixed at a specific location. Additionally, static and dynamic obstacles can be classified based on the LiDAR sensor into opaque obstacles that reflect lasers and transparent obstacles that transmit lasers. For example, a transparent obstacle could be a glass door, while a dynamic obstacle could be a person or human footprints.

[0084] In various embodiments, each coordinate of the point cloud acquired by the processor (150) corresponds to an indoor space, and free space within the indoor space may be excluded. Typically, when a laser emitted from a LiDAR sensor passes through, a conventional map generation device defines this as free space. However, since transparent obstacles such as glass doors also allow lasers to pass through, an error may occur in an indoor map generation device (100) containing only a 2D LiDAR sensor (140) where a location where a transparent obstacle is placed is defined as free space. Accordingly, the processor (150) can recognize only obstacles excluding free space, without recognizing both obstacles and free space using the 2D LiDAR sensor (140). For example, the processor (150) can change the setting value for recognizing free space among the map generation parameters set in the SLAM algorithm so that the 2D LiDAR sensor (140) collects only the laser reflected from obstacles.

[0085] In various embodiments, the processor (150) can identify the type of sensor included in the device. Specifically, at the beginning of the indoor space exploration, the processor (150) can identify whether the sensor included in the device is a single LiDAR sensor, and the angular resolution and scan frequency value of the LiDAR sensor. Accordingly, if the device includes only a low-spec LiDAR sensor, the processor (150) can adjust the map generation parameters of the device. For example, the processor (150) can change a parameter set so that any one point can be recognized as free space from 'true' to 'false'.

[0086] In this way, the processor (150) can obtain a point cloud corresponding to an indoor space containing a transparent obstacle by adjusting map generation parameters. The point cloud here may include points obtained near the angle of incidence perpendicular to the glass surface. In other words, the processor (150) can obtain points to distinguish it as an obstacle without recognizing the transparent obstacle as free space.

[0087] Meanwhile, the processor (150) can accumulate points for an indoor space to obtain a point cloud, thereby generating a cumulative occupancy grid map. In the cumulative occupancy grid map generated by the processor (150) through a SLAM algorithm, each grid may contain a probability that an obstacle exists.

[0088] After step S110, the processor (150) can cluster the point cloud to generate an occupancy grid map for an indoor space containing static obstacles (S120). Specifically, as the processor (150) collects all points through the LiDAR sensor, the point cloud may contain dynamic (or transient) obstacles as well as transparent obstacles. Accordingly, the processor (150) can extract discontinuous areas from the point cloud to remove dynamic obstacles.

[0089] In this regard, FIG. 5 is an exemplary diagram illustrating a point cloud collected by an indoor map generating device according to one embodiment of the present invention.

[0090] Referring to FIG. 5, when generating an indoor space map for a corridor, an initial occupancy grid map such as (a) can be generated based on the point cloud obtained through step S110. The initial grid map may contain dynamic obstacles in discontinuous areas as shown in (b). The processor (150) can extract areas where discontinuous points exist and cluster the areas to remove clusters smaller than a preset size. Accordingly, the processor (150) can generate a grid map with dynamic obstacles removed as shown in (c). Through this process, the processor (150) can generate a grid map in which discontinuous areas extracted by glass doors are removed and discontinuous areas extracted by people are removed.

[0091] After step S120, the processor (150) can acquire point clouds for the indoor space multiple times to resolve obstacle recognition errors through clustering and update points matched to the occupancy grid map (Update Occupancy Grid Mapping) (S130). Specifically, the indoor map generating device (100) can acquire point clouds a predetermined number of times at predetermined time intervals and accumulate points in the point clouds. The processor (150) can autonomously move within the indoor space based on the device's starting position, ending position, and direction data provided in advance by the user, thereby eliminating unnecessary tasks that require an administrator to monitor the indoor map generating device (100) in real time. The processor (150) can acquire, for example, N point clouds (N is a natural number greater than or equal to 2) through autonomous movement, and can acquire the first to Nth point clouds through point accumulation. The processor (150) can distinguish between static obstacles and dynamic obstacles in the accumulated point clouds. In other words, the processor (150) can calibrate the occupancy grid map (Occupancy Grid Gap Calibration via Navigation) through the device's own navigation without the control of a manager.

[0092] In various embodiments, the processor (150) can calculate the probability that a static obstacle exists at each point constituting the accumulated point cloud. That is, the processor (150) can perform probabilistic occupancy grid mapping through the point clouds that are repeatedly acquired. For example, the processor (150) can calculate the probability that a static obstacle exists for each grid in an occupancy grid map based on each of the first to Nth point clouds. The probability of a static obstacle existing is determined by the sum of the probability values ​​in each point cloud, and the processor (150) can determine the weight for each of the N point clouds (i.e., the first to Nth point clouds) acquired a preset number of times. For example, the processor (150) can apply a weight of 1 / (1+M) to each of the N point clouds according to the number of times the point clouds are acquired (M, where M is a natural number greater than or equal to 2). The processor (150) can multiply the probability values ​​existing for each grid in each of the N point clouds by a weight and can sum the multiplication results calculated for each grid. The processor (150) can determine whether this sum value is a preset value. If the sum value is greater than or equal to the preset value, the processor (150) can determine that a static obstacle exists in the corresponding grid.

[0093] Meanwhile, the value used as a criterion for determining the presence of static obstacles may vary depending on the environment. For example, since the probability of the existence of dynamic and static obstacles differs between environments with few dynamic obstacles (e.g., unmanned factories, automated warehouses, logistics centers, etc.) and environments with few dynamic obstacles (e.g., unmanned factories, automated warehouses, logistics centers, etc.), the processor (150) may receive a reference value from a user or adjust the reference value by evaluating the environment of dynamic obstacles in an indoor space (e.g., evaluating the environment based on the ratio of continuous and discontinuous areas within a point cloud). For example, the processor (150) may lower the reference value in environments with few dynamic obstacles (Low Threshold for Obstacles) and raise the reference value in environments with many dynamic obstacles (High threshold for Obstacles). That is, the processor (150) can generate a map using a reference value appropriate to the environment in each of the environments with many dynamic obstacles and environments with few dynamic obstacles.

[0094] Up to now, an indoor map generating device (100) according to one embodiment of the present invention has been described. According to the present invention, not only can obstacles having transparent properties be distinguished even with low-cost hardware, but static obstacles and dynamic obstacles can also be clearly distinguished, thereby securing cost competitiveness in large-scale industrial environments.

[0095] Below, a series of methods for generating an indoor map with precision through iterative updates of the map in an indoor map generation system (1000) will be described.

[0096] FIG. 6 is a detailed flowchart of an indoor map generation method according to one embodiment of the present invention.

[0097] Referring to FIG. 6, the indoor map generating device (100) can optimize parameters according to the type of sensor included in the device. For example, if the indoor map generating device (100) includes only one LiDAR sensor, it can optimize parameters so as not to recognize free space. In STEP 1, the indoor map generating device (100) can generate a cumulative grid map containing the probability of the existence of obstacles through driving in the indoor space. Here, the grid map may be a grid map that recognizes only obstacles, excluding the recognition of free space. In STEP 2, the indoor map generating device (100) can remove clusters smaller than a preset size in discontinuous areas through clustering. That is, the indoor map generating device (100) can generate a grid map with obstacles removed through cluster removal. In STEP 3, the indoor map generating device (100) drives autonomously in the indoor space and can match the grid map with obstacles removed with the grid map obtained through repeated autonomous driving thereafter. Here, matching can be understood as updating the probability of an obstacle existing in each grid of the accumulated point cloud. In STEP 4, the indoor map generating device (100) maps the grid map that has not performed clustering with the map updated in STEP 3 to finally determine whether an obstacle exists in each grid. The indoor map generating device (100) can determine a reference value for judgment based on the environment of the indoor space. Based on the determined reference value, the indoor map generating device (100) can determine that there is no static obstacle if the probability value is lower than the reference value, and conversely, determine that there is a static obstacle if the probability value is higher than the reference value.

[0098] FIG. 7 is an exemplary diagram illustrating a map produced by an indoor map generating device according to one embodiment of the present invention.

[0099] Referring to FIG. 7, when only the SLAM algorithm is performed using points acquired from a conventional low-cost, low-spec LiDAR sensor, a static obstacle having transparent properties is not recognized as in (a). However, when using the indoor map generation method according to one embodiment of the present invention, the indoor map generation method defined in FIG. 6, a map in which a static obstacle having transparent properties is recognized can be generated as in (b).

[0100] Up to now, an indoor map generation method according to an embodiment of the present invention has been described. According to the present invention, by accumulating point clouds with parameters adjusted, obstacles having transparent properties can be accurately recognized using only a LiDAR sensor. Furthermore, obstacles can be accurately recognized even in situations with many variables by calculating probability values ​​of a grid map generated from the accumulated point clouds. For example, even if points are acquired in situations where a transparent door opens or closes, a map in which static obstacles having transparent properties are distinguished can be generated by summing the probability values.

[0101] Although embodiments of the present invention have been described in more detail with reference to the attached drawings, the present invention is not necessarily limited to these embodiments and may be modified in various ways within the scope of the technical spirit of the present invention. Accordingly, the embodiments disclosed in the present invention are intended to explain, not limit, the technical spirit of the present invention, and the scope of the technical spirit of the present invention is not limited by these embodiments. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. The scope of protection of the present invention shall be interpreted by the claims below, and all technical spirits within an equivalent scope shall be interpreted as being included within the scope of rights of the present invention.

Claims

1. A map generation method performed by a processor of an indoor map generation device, A step of acquiring a point cloud corresponding to an indoor space containing static and dynamic obstacles; A step of clustering the above point cloud to generate an occupancy grid map for an indoor space including the above static obstacles; and A method for generating an indoor map comprising the step of acquiring a point cloud corresponding to the indoor space multiple times and updating the points matched to the occupancy grid map.

2. In Paragraph 1, The step of acquiring the above point cloud is, A method for generating an indoor map, comprising the step of acquiring a point cloud corresponding to the indoor space, excluding the free space within the indoor space.

3. In Paragraph 2, The step of acquiring the above point cloud is, A step of identifying the type of sensor included in the above device, and An indoor map generation method further comprising the step of adjusting map generation parameters of the device according to the type of sensor.

4. In Paragraph 3, The step of acquiring the above point cloud is, A method for generating an indoor map, comprising the step of obtaining a point cloud corresponding to an indoor space that includes obstacles having transparent attributes among the static obstacles, based on the adjustment of the map generation parameters.

5. In Paragraph 1, The step of generating the above-mentioned occupancy grid map is, The step of extracting discontinuous regions from the above point cloud, and An indoor map generation method further comprising the step of clustering the above discontinuous regions to remove clusters smaller than a preset size.

6. In Paragraph 1, The step of updating the above points is, A step of acquiring a point cloud a predetermined number of times at a predetermined time interval and accumulating points in the point cloud, and An indoor map generation method further comprising the step of distinguishing between the static obstacle and the dynamic obstacle in an accumulated point cloud.

7. In Paragraph 6, The step of distinguishing between the static obstacle and the dynamic obstacle is, An indoor map generation method further comprising the step of calculating the probability that a static obstacle exists at each point constituting the accumulated point cloud.

8. In Paragraph 7, The step of calculating the above probability is, A step of determining the weight of each of the N point clouds acquired a predetermined number of times, and An indoor map generation method further comprising the step of determining whether the sum of the weights of each of the N point clouds is greater than or equal to a preset value.

9. Communication interface; Memory; and A processor operably connected to the communication interface and the memory; comprising The above processor is, An indoor map generation device configured to acquire a point cloud corresponding to an indoor space containing static obstacles and dynamic obstacles, cluster the point cloud to generate an occupancy grid map for the indoor space containing the static obstacles, acquire the point cloud corresponding to the indoor space multiple times, and update points matched to the occupancy grid map.

10. In Paragraph 9, The above processor is, An indoor map generating device configured to acquire a point cloud corresponding to the indoor space, excluding the free space within the indoor space.

11. In Paragraph 10, The above processor is, An indoor map generation device further configured to identify the type of sensor included in the device to acquire the above point cloud, and to adjust the map generation parameters of the device according to the type of sensor.

12. In Paragraph 11, The above processor is, An indoor map generation device configured to acquire a point cloud corresponding to an indoor space, including obstacles having transparent attributes among the static obstacles, based on the adjustment of the map generation parameters above.

13. In Paragraph 10, The above processor is, An indoor map generating device further configured to generate the above-mentioned occupancy grid map by extracting discontinuous regions from the point cloud and clustering the discontinuous regions to remove clusters smaller than a preset size.

14. In Paragraph 10, The above processor is, An indoor map generating device further configured to acquire a point cloud a predetermined number of times at a predetermined time interval to update the above points, accumulate points in the point cloud, and distinguish between the static obstacle and the dynamic obstacle in the accumulated point cloud.

15. In Paragraph 14, The above processor is, An indoor map generating device further configured to calculate the probability that a static obstacle exists at each point constituting the accumulated point cloud in order to distinguish between the static obstacle and the dynamic obstacle.

16. In Paragraph 15, The above processor is, An indoor map generating device further configured to determine the weight of each of the N point clouds obtained a preset number of times to calculate the above probability, and to determine whether the sum of the weights of each of the N point clouds is greater than or equal to a preset value.

Images