Method and system for simultaneous localization and mapping based on 2D lidar and camera

Abstract

A method and system for simultaneous 2D lidar and camera-based localization and mapping, comprising loading a map to which 2D feature point data is bound, acquiring 2D LiDAR point data from a LiDAR (10) attached to a mobility (60), acquiring 3D observation feature point data by receiving a front image from a camera (40) attached to the mobility (60), reducing a dimension of the 3D observation feature point data to 2D observation feature point data, binding the 2D observation feature point data to the 2D LiDAR point data, and acquiring an absolute position of the mobility (60) based on the 2D observation feature point data bound to the 2D LiDAR point data and the map to which the 2D feature point data is bound.

Description

TECHNICAL AREA

[0001] The present disclosure / invention relates to a method and system for simultaneous localization and mapping (SLAM, e.g., simultaneous localization and mapping). BACKGROUND

[0002] Simultaneous localization and mapping (SLAM) can refer to a technology that simultaneously detects the current location of a mobility (e.g., a mobile object, e.g., a mobile device, e.g., a vehicle) and creates a map of the surroundings while exploring an unknown environment.

[0003] A state-of-the-art autonomous driving mobility-based indoor SLAM acquires spatial information using a stereo image or a light detection and ranging device (LiDAR), and then calculates the motion information using a position detection sensor such as a value encoder (e.g., a rotary encoder) or an inertial sensor to create a map. However, a spatial information generation sensor such as an image or LiDAR has the disadvantage of being sensitive to illuminance and lacking sufficient resolution, so the sensor may generate incorrect maps. This may lead to a mobility collision accident.

[0004] For example, if a chair or table with thin legs is placed indoors, even though the LiDAR detects the thin legs, they are treated as noise, which is largely eliminated by correction during a mapping process. Because a topographical feature is lost, it is difficult to perform SLAM in indoor spaces with a variety of similar spatial structures, such as offices.

[0005] The matters described in this Background section are provided to provide a better understanding of the background of the present disclosure / invention and may also relate to matters that are not prior art and that are not already publicly known, in use, or available. BRIEF EXPLANATION OF THE INVENTION

[0006] The present disclosure / invention relates to a method and system for simultaneous localization and mapping (SLAM), and more particularly may relate to a method and system for simultaneous localization and mapping that connects (e.g., links) an observation feature point (e.g., observation feature point) observed (e.g., captured) by a camera to a two-dimensional (2D) LiDAR point and compares the observation feature point bound to the 2D LiDAR point (e.g., linked to the 2D LiDAR point) to a map to which a feature point is bound (e.g., linked to a feature point) to accurately locate the mobility (e.g., the mobile object, e.g., the mobile device, e.g., the vehicle) and create a map based on the mobility localization (e.g., localizing the mobility).

[0007] Some embodiments of the present disclosure / invention may provide a simultaneous localization and mapping (SLAM) method and system that connects an observation feature point observed by a camera to a 2D LiDAR point and compares the observation feature point bound to the 2D LiDAR point to a map to which a feature point is bound to accurately locate mobility and create a map based on the mobility localization.

[0008] According to an exemplary embodiment of the present disclosure / invention, a method for simultaneous localization and mapping (e.g., a simultaneous localization and mapping method) comprises: loading, by a control unit, a map to which 2D feature point data stored in a memory is bound; acquiring, by the control unit, 2D LiDAR point data from a LiDAR deployed in a mobility (e.g., a LiDAR equipped with a mobility); acquiring, by the control unit, three-dimensional (3D) observation feature point data by receiving, by the control unit, a front image of the mobility (e.g., a forward-looking image as viewed from the mobility) from a camera deployed in the mobility; reducing, by the control unit, a dimension of the 3D observation feature point data to 2D observation feature point data;Connecting (e.g., linking) the 2D observation feature point data to the 2D LiDAR point data by the control unit, and detecting (e.g., detecting, e.g., determining) an absolute position of the mobility based on the 2D observation feature point data linked to the 2D LiDAR point data and (e.g., based on) the map to which the 2D feature point data is linked, by the control unit.

[0009] Detecting an absolute position of the mobility based on the 2D observation feature point data bound to the 2D LiDAR point data and the map to which the 2D feature point data is bound may further comprise selecting a calculated position of the mobility, comparing the bound observation feature point data with the feature point data bound to the calculated position of the map, and selecting a final position of the mobility.

[0010] The final mobility position can be selected by checking a similarity of the bound observation feature point data and the feature point data bound to the calculated map position.

[0011] The calculated position of the mobility can be selected by localizing the mobility.

[0012] The method for simultaneous localization and mapping may further comprise updating, by the control unit, the map to which the 2D feature point data is bound based on the absolute position of the mobility and the 2D observation feature point data bound to the 2D LiDAR point data.

[0013] The method for simultaneous localization and mapping may further comprise storing, by the control unit, the updated map to which the 2D feature point data is bound in memory.

[0014] The method for simultaneous localization and mapping may further comprise, before reducing a dimension of the 3D observation feature point data to 2D observation feature point data, deleting collision-free (e.g., non-colliding) 3D observation feature point data that does not collide with mobility from the 3D observation feature point data by the control unit.

[0015] Mobility specifications can be stored in memory and the collision-free 3D observation feature point data can be selected based on the mobility specifications.

[0016] According to an exemplary embodiment of the present disclosure / invention, a system for simultaneous localization and mapping (e.g., a simultaneous localization and mapping system) comprises a LiDAR mounted in a mobility device and configured to acquire 2D LiDAR point data, a camera mounted in (e.g., on) the mobility device and configured to acquire a forward image of the mobility device, and a controller (e.g., a controller) configured to: load a map stored in memory to which 2D feature point data is bound; receive the 2D LiDAR point data from the LiDAR device; receive the image from the camera; acquire 3D observation feature point data from the image (e.g., based on the image); reduce a dimension of the 3D observation feature point data to 2D observation feature point data;- to bind the 2D observation feature point data to the 2D LiDAR point data and - to capture an absolute position of the mobility based on the 2D observation feature point data bound to the 2D LiDAR point data and the (for example, based on) the map to which the 2D feature point data is bound.

[0017] The control unit may be further configured to - select a calculated mobility position, - compare the bound observation feature point data with the feature point data bound to the calculated map position, and - select a final mobility position based on the comparison result.

[0018] The control unit may be further configured to select the final position of the mobility by checking a similarity of the bound observation feature point data and the feature point data bound to the calculated position of the map.

[0019] The simultaneous localization and mapping system may further include a motion data sensor mounted in the mobility and configured to collect motion data of the mobility, and the controller may be configured to select the calculated position of the mobility by localizing the mobility based on the 2D LiDAR point data or the motion data of the mobility (e.g., the mobility motion data).

[0020] The controller may be further configured to update the map to which the 2D feature point data is bound based on the absolute position of the mobility and the 2D observation feature point data bound to the 2D LiDAR point data.

[0021] The controller may be further configured to store the updated map to which the 2D feature point data is bound in memory.

[0022] The control unit may be further configured to delete collision-free 3D observation feature point data that does not collide with mobility from the 3D observation feature point data before reducing the dimension of the 3D observation feature point data to the 2D observation feature point data.

[0023] Mobility specifications may be stored (e.g., stored) in memory, and the controller may be configured to select the collision-free 3D observation feature point data based on the mobility specifications.

[0024] According to some embodiments of the present disclosure / invention, an observation feature point observed by a camera may be bound to a 2D LiDAR point and may be compared to a map to which a feature point is bound to accurately locate the mobility. Furthermore, the map may be accurately created (e.g., generated, e.g., updated) based on the accurate location of the mobility.

[0025] Furthermore, perhaps unlike a SLAM that uses an artificial neural network, the 2D LiDAR point and the observation feature point observed by the camera can be bound (e.g., linked), allowing precise localization without complex calculations. Accordingly, a control board with advanced computing capabilities can be dispensed with, and a central processing unit (CPU) integrated into the existing mobility can be used.

[0026] Furthermore, advantages and functionalities that may be achieved or predicted by some embodiments of the present disclosure / invention are disclosed directly or implicitly in the detailed description of some example embodiments of the present disclosure / invention. That is, various advantages and functionalities may be predicted according to some embodiments of the present disclosure / invention and are disclosed in the detailed description below. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Exemplary embodiments of the present disclosure / invention may be better understood by referring to the following description in conjunction with the accompanying drawings, in which like reference numerals may designate identical or functionally similar elements, and in which: Fig. 1 is a block diagram of a system for simultaneous localization and mapping according to an embodiment of the present disclosure / invention, Fig. 2 is a flowchart of a method for simultaneous localization and mapping according to an embodiment of the present disclosure / invention, Fig. 3 is a diagram illustrating a 3D observation feature point captured by a camera, according to some embodiments of the present disclosure / invention, Fig. 4 is a diagram illustrating 2D observation feature point data bound to 2D LiDAR point data, according to some embodiments of the present disclosure / invention Fig. 5 is a diagram illustrating a calculated position of a mobility, according to some embodiments of the present disclosure / invention Fig. 6 is a diagram illustrating a result obtained by comparing bound observation feature point data and feature point data bound to a calculated position in a map, according to some embodiments of the present disclosure / invention, and Fig. 7 is a diagram illustrating updating a map to which a feature point is bound, according to some embodiments of the present disclosure / invention.

[0028] It should be understood that the above-mentioned drawings are not necessarily drawn to scale, and somewhat simplified representations of various features illustrate the basic principles of some embodiments of the present disclosure / invention. For example, certain design features of some embodiments of the present disclosure may have a particular dimension, direction, position, and shape, which may be determined in part by a specific intended application and usage environment. DETAILED DESCRIPTION OF EMBODIMENTS

[0029] The singular forms used herein may also include / comprise plural forms unless the context clearly indicates otherwise. The terms "include" / "comprise" and / or "including" / "comprising," when used herein, denote the presence of the recited features, integers, steps, operations, ingredients, and / or components, but these terms are also to be understood as not excluding the presence or addition of one or more other features, integers, steps, operations, ingredients, components, and / or groups thereof. As used herein, the term "and / or" includes any or all combinations of the related listed elements.

[0030] A “mobility” or other similar terms used in the description may include general land mobility (e.g., a device for moving on land) such as passenger vehicles, including sport utility vehicles (SUVs), buses, trucks, tractors, and various commercial vehicles; maritime mobility (e.g., a device for moving on / in water), including various boats and ships; and air mobility (e.g., a device for moving in the air, including airplanes and drones), as well as any object that can move by consuming a power source, such as robots or motorized wheelchairs and other mobility devices that often move indoors.Furthermore, "mobility" or other similar terms used in the description may include, for example, hybrid mobility, electric mobility, plug-in hybrid mobility, hydrogen mobility, and mobility with other alternative fuels (e.g., fuels derived from sources other than petroleum). As mentioned in the description, hybrid mobility may include mobility with two or more energy sources, e.g., gasoline-powered and electric-powered mobility. Mobility according to some embodiments of the present disclosure may include not only manually powered mobility, but also remotely controlled, autonomous, and / or automatically powered mobility.

[0031] In addition, one or more of the following methods and embodiments may be performed by at least one or more control units. The term "control unit" may refer to a hardware device that includes at least one memory and at least one processor. The memory may be configured to store code and program instructions, and the processor may be particularly programmed to execute code and program instructions to perform one or more processes or operations, for example, as described in more detail below. The control unit may control the operation of units, modules, components, devices, appliances, or the like, as described herein. Furthermore, the following methods may be performed by a device that includes a control unit along with one or more other components, as would be known to a person of ordinary skill in the art.

[0032] Furthermore, a control unit of the present disclosure / invention may be implemented as a non-transitory, computer-readable recording medium containing executable program instructions executed by at least one processor. Examples of computer-readable recording media may include, but are not necessarily limited to, ROM, RAM, compact disk (CD) ROM, magnetic tapes, floppy disks, flash drives, smart cards, and optical data storage devices. The computer-readable recording medium may also be distributed across a computer network so that program instructions can be stored and executed in a distributed manner, for example, on a telematics server or a controller area network (CAN).

[0033] In the following, exemplary embodiments of the present disclosure will be described in more detail with reference to the accompanying drawings.

[0034] Fig. 1 is a block diagram of a system for simultaneous localization and mapping according to an embodiment of the present disclosure / invention.

[0035] As in Fig. 1, a simultaneous localization and mapping (SLAM) system according to an embodiment of the present disclosure / invention may include a LiDAR 10, a transducer (e.g., encoder) 20, an inertial sensor 30, a camera 40, a controller 50, and a mobility 60.

[0036] The LiDAR 10 may be mounted in the mobility 60 and record the time after / within which a laser pulse reflected from an object (e.g., a fixed topography and an obstacle) within a detection range of the LiDAR 10 returns after the laser pulse was emitted in front of the mobility 60 to record information about the object, such as a distance between the LiDAR 10 and the object, as well as a direction, a speed, a temperature, a material distribution, and / or a concentration property of the object. The object may be another mobility, people, objects, pillars, or walls located outside the mobility 60 equipped with the LiDAR 10 as a sensor, but an embodiment of the present disclosure / invention is not necessarily specifically limited to the type of object.The LiDAR 10 may be connected to the control unit 50 to acquire 2D LiDAR point data (e.g., 2D data of a plurality of LiDAR points) within the detection range and transmit the 2D LiDAR point data to the control unit 50.

[0037] The sensor 20 can detect information about the rotation of a drive motor (e.g., an electric drive motor) or a wheel installed in the mobility device 60. The sensor 20 can be connected to the control unit 50 to transmit the detected information about the rotation of the drive motor or wheel to the control unit 50. The control unit 50 can calculate the movement data of the mobility device 60, such as a movement speed and / or a movement distance of the mobility device 60, based on the information about the rotation of the drive motor or wheel.

[0038] The inertial sensor 30 can detect information about a movement situation of the mobility 60, including a speed, a direction, a gravity, and an acceleration of the mobility 60. The inertial sensor 30 can be connected to the control unit 50 to transmit the detected information about the movement situation of the mobility 60 to the control unit 50. The control unit 50 can detect or supplement the movement data of the mobility 60 based on the information about the movement situation of the mobility 60.

[0039] Both the encoder 20 and the inertial sensor 30 may be used as a motion data sensor that acquires the motion data of the mobility 60, or only one of the encoder 20 and the inertial sensor 30 may be used as a motion data sensor. Furthermore, the motion data sensor is not necessarily limited to the encoder 20 and the inertial sensor 30, but may include various sensors that acquire the motion data of the mobility 60.

[0040] The camera 40 may be installed in the mobility device 60 and capture a front image of the mobility device 60 within a detection range of the camera 40. The camera 40 may be connected to the control unit 50 to transmit the captured image to the control unit 50. The image may be configured by pixel data with a plurality of pixels. A size of the pixel data of the image may be (N*M), and N and M are natural numbers.

[0041] The control unit 50 can receive the 2D LiDAR point data from the LiDAR 10, receive the information about the rotation of the drive motor or the wheel from the encoder 20, receive the information about the movement situation of the mobility 60 from the inertial sensor 30, and receive the front image of the mobility 60 from the camera 40.

[0042] The control unit 50 may be configured to - detect an object (e.g., a 3D observation feature point) in the image based on the received image using an object detection algorithm to acquire 3D observation feature point data (e.g., an object detection algorithm to acquire 3D observation feature point data), - delete collision-free 3D observation feature point data that do not collide with the mobility 60 from the 3D observation feature point data, - reduce one dimension of the 3D observation feature point data to 2D observation feature point data using a known dimension reduction algorithm, and - link the 2D observation feature point data to (e.g., with) a factor graph using a known binding algorithm (e.g., linking algorithm). surrounding 2D LiDAR point data.

[0043] In one example, the control unit 50 may acquire mobility movement data based on the information about the rotation of the drive motor or wheel received from the encoder 20 or the information about the movement situation of the mobility 60 received from the inertial sensor 30, and select calculated positions of the mobility 60 based on the mobility movement data. In another example, the control unit 50 may select the calculated positions of the mobility 60 based on the 2D LiDAR point data using a known localization algorithm (e.g., AMCL). To select the calculated positions of the mobility 60, a method for positioning the mobility 60 other than the method described here may be used.

[0044] The control unit 50 may include a memory 52, and a map to which the 2D feature point data is bound may be stored in the memory 52. ​​In one example according to an embodiment of the present disclosure / invention, the map to which the 2D feature point data is bound may be a map created by the SLAM method that was / is previously performed, and the map may include the 2D feature point data bound to the 2D LiDAR point data. However, the map to which the 2D feature point data is bound is not necessarily limited to the map created by the SLAM method according to the embodiment of the present disclosure / invention, and as long as the 2D feature point data bound to the 2D LiDAR point data is included, a mapping method is not specifically limited.In order to distinguish the feature point data currently observed by the camera 40 from the feature point data included in the map, the feature point data currently observed by the camera 40 may be referred to as observation feature point data.

[0045] The control unit 50 may compare the 2D observation feature point data bound to the 2D LiDAR point data with the 2D feature point data bound to the calculated map position. The control unit 50 may compare the 2D observation feature point data bound to the 2D LiDAR point data with the 2D feature point data bound to the calculated map position to check similarity using a known feature point similarity checking algorithm. Thereafter, the control unit 50 may select (e.g., determine, e.g., set) the calculated position with the highest similarity as the final position of the mobility 60 and acquire an absolute position of the mobility 60, for example, through a global localization method.The control unit 50 may update the map to which the 2D feature point data is bound based on the absolute position of the localized mobility 60 and the 2D observation point data bound to the 2D LiDAR point data, and store the updated map in the memory 52.

[0046] The control unit 50 may include one or more microprocessors, and the one or more microprocessors may be programmed to perform operations of a SLAM method according to some embodiments of the present disclosure / invention.

[0047] According to some embodiments of the present disclosure, the control unit 50 may be connected to the mobility 60 and establish a route (e.g., a movement path) of the mobility 60 or control the movement of the mobility 60 using the map updated by the SLAM method. For example, the control unit 50 may control the mobility 60 to follow the object (e.g., move toward the object) or control the mobility 60 to avoid the object.

[0048] Fig. 2 is a flowchart of a method for simultaneous localization and mapping according to an embodiment of the present disclosure / invention. Fig. 3 illustrates a 3D observation point captured by a camera in an example. Fig. Figure 4 illustrates 2D observation feature point data bound to 2D LiDAR point data in an example. Fig. 5 illustrates a calculated position of a mobility in an example. Fig. 6 shows a result obtained by comparing bound observation point data and feature point data bound to a calculated position in a map in an example. Fig. 7 shows an example of updating a map to which a feature point is bound.

[0049] As in the Fig. 2 to 7, a SLAM method according to an embodiment of the present disclosure / invention may begin with loading the map 80 to which the 2D feature point data 84 is bound, in operation S110. In one example, the map 80 to which the 2D feature point data 84 is bound is the map created by the SLAM method according to some embodiments of the present disclosure / invention, which was previously performed, and, for example, as shown in Fig. 6, the map 80 contains the 2D feature point data 84 bound to the 2D LiDAR point data (see lines connecting adjacent circles). The map 80 to which the 2D feature point data 84 is bound may be stored in the memory 52 of the control unit 50.

[0050] The LiDAR 10 may be installed in the mobility device 60, may acquire (e.g., acquire) the 2D LiDAR point data 72 in operation (e.g., during the operation, e.g., during step S120), and may transmit the acquired 2D LiDAR point data 72 to the control unit 50. Furthermore, the camera 40 may be installed in the mobility device 60, acquire the front image of the mobility device 60 (or the image of the environment in front of the mobility device 60), and transmit the acquired image to the control unit 50. The control unit 50 may acquire the 3D observation feature point data 70 in the image based on the received image using an object detection algorithm in operation S130. Fig. 3 shows an example illustrating the 3D observation feature point data 70 in the image captured by the camera 40.

[0051] When the 3D observation feature point data 70 is / has been acquired, the control unit 50 may delete the collision-free 3D observation feature point data that does not collide with the mobility 60 from the 3D observation feature point data 70 in operation S140. For example, the feature points above the highest point of the mobility 60 may be the collision-free feature points where a collision with the mobility 60 is unlikely. Accordingly, the collision-free 3D observation feature point data where a collision with the mobility 60 is unlikely may be removed from the 3D observation feature point data 70 to reduce computational complexity. Specifications of the mobility 60 (e.g.,A specific specification (e.g., a diameter of the wheel or a maximum height of the mobility 60) may be stored in the memory 52, and the control unit 50 may select the collision-free 3D observation feature point data based on the specifications of the mobility 60. Operation S140 may be an optional operation of a particular method according to some embodiments of the present disclosure / invention.

[0052] The control unit 50 may reduce the dimension of the 3D observation feature point data 70 from which the collision-free 3D observation feature point data is / has been deleted to the 2D observation feature point data 70 using a dimension reduction algorithm in operation S150. The dimension reduction algorithm that reduces one dimension of the 3D data to the 2D data may be obvious (e.g., known) to a person with ordinary knowledge in the art, and accordingly, a detailed description may be omitted.

[0053] The control unit 50 may connect the 2D observation feature point data 70 with the surrounding 2D LiDAR point data 72 using a binding algorithm (e.g., a factor graph) in operation S160. Fig. 4 shows examples of 2D observation feature point data 70 bound to at least some of the 2D LiDAR point data 72. For example, each of five observation feature point data 70 may be bound to two or more 2D LiDAR point data 72. In one example, using the binding algorithm, the feature point may be bound to the LiDAR point based on a distance and / or direction between the feature point and the LiDAR point. Such a binding algorithm may be obvious (e.g., known) to a person of ordinary skill in the art, and therefore, a detailed description may be omitted.

[0054] The control unit 50 may select the calculated positions 82a, 82b, 82c, and 82d of the mobility 60 through a positioning method of the mobility 60 in operation S170. The positioning method using the 2D LiDAR 10, the encoder 20, and / or the inertial sensor 30 may include an error in the indoor environment. To reduce the error, in some embodiments, the control unit 50 does not confirm the position of the mobility 60 detected with the 2D LiDAR 10, the encoder 20, and / or the inertial sensor 30, but may instead select the calculated positions that were calculated as where the mobility 60 is located. Fig. 5 shows that the first, second, third and fourth calculated positions 82a, 82b, 82c and 82d can be selected.

[0055] After the calculated position of the mobility 60 has been / is selected, the control unit 50 may compare the 2D observation feature point data 70 bound to the 2D LiDAR point data 72 with the 2D feature point data 84 bound to the calculated positions 82a, 82b, 82c, and 82d on the map 80 in operation S180. The control unit 50 may check (e.g., verify) the similarity of the 2D observation feature point data 70 bound to the 2D LiDAR point data 72 with the 2D feature point data 84 bound to the calculated positions 82a, 82b, 82c, and 82d on the map 80 using an algorithm for checking similarity by a shape and a position of the feature point, and may select the calculated position with the highest similarity as the final position of the mobility 60 in operation S190.

[0056] As in Fig. 6, for example, a similarity between the 2D observation feature point data 70 based on the first calculated position 82a and the 2D feature point data 84 is 20%, a similarity between the 2D observation feature point data 70 based on the second calculated position 82b and the 2D feature point data 84 is 40%, a similarity between the 2D observation feature point data 70 based on the third calculated position 82c and the 2D feature point data 84 is 10%, and a similarity between the 2D observation feature point data 70 based on the fourth calculated position 82d and the 2D feature point data 84 is 90%. Accordingly, the fourth calculated position 82d having the highest similarity can be selected as the final position of the mobility 60.

[0057] Thereafter, the control unit 50 may determine the absolute position of the mobility 60 based on the selected final position using a global localization method in operation S200. As in Fig. 7, the control unit 50 may determine that the mobility 60 is located in the fourth calculated position 82d and acquire the position of the mobility 60 based on the 2D observation feature point data 70 bound to the 2D LiDAR point data 72 using a global localization method.

[0058] The control unit 50 may update the map 80 to which the 2D feature point data 84 is bound based on the absolute position of the mobility 60 and the 2D observation feature point data 70 bound to the 2D LiDAR point data 72, and may store the updated map 80 in the memory 52. ​​In the updated map 80, the 2D observation feature point data 70 bound to the 2D LiDAR point data 72 may be stored as the 2D feature point data 84 bound to the 2D LiDAR point data 72.

[0059] Although some embodiments of the present disclosure / invention have been described in connection with exemplary embodiments presently considered practical, the present disclosure / invention is not necessarily limited to the disclosed exemplary embodiments. On the contrary, the present disclosure / invention is intended to cover various modifications and equivalent arrangements falling within the scope of the appended claims, with the present invention being defined by the appended claims.

Claims

[1] A method for localization and mapping, comprising: Loading (S110) a map to which two-dimensional (2D) feature point data are bound, Acquiring (S120) 2D LiDAR point data from a light detection and ranging device (LiDAR) (10) with which a mobility (60) is equipped, Acquiring three-dimensional (3D) observation feature point data by receiving a front image from a camera (40) with which the mobility is equipped, Reducing (S150) one dimension of the 3D observation feature point data to 2D observation feature point data, Binding (S160) the 2D observation feature point data to the 2D LiDAR point data, and Detecting an absolute position of the mobility (60) based on the 2D observation feature point data bound to the 2D LiDAR point data and based on the map to which the 2D feature point data is bound. [2] The method of claim 1, wherein detecting the absolute position of the mobility (60) comprises: Selecting (S170) a calculated position of mobility (60), Comparing (S180) the bound 2D observation feature point data with the 2D feature point data bound to the calculated position of the map, and Selecting (S190) a final mobility position based on the comparison. [3] The method of claim 2, wherein selecting (S190) the final position of the mobility comprises checking a similarity of the bound 2D observation feature point data and the 2D feature point data bound to the calculated position of the map. [4] The method of claim 2 or 3, wherein selecting (S180) the calculated position of the mobility (60) comprises locating the mobility (60). [5] The method of any one of claims 1 to 4, further comprising updating the map to which the 2D feature point data is bound based on the absolute position of the mobility (60) and based on the 2D observation feature point data bound to the 2D LiDAR point data. [6] The method of claim 5, further comprising storing the updated map in a memory (52). [7] The method according to any one of claims 1 to 6, wherein prior to reducing (S150) the dimension of the 3D observation feature point data to 2D observation feature point data, collision-free 3D observation feature point data that do not collide with the mobility (60) are deleted (S140) from the 3D observation feature point data. [8] The method of claim 7, wherein specifications of mobility (60) are stored in a memory (52), and the method further comprises selecting the collision-free 3D observation feature point data based on the specifications of mobility. [9] Localization and mapping system comprising: a light detection and distance measuring device (LiDAR) (10) with which a mobility (60) is equipped and which is configured to capture two-dimensional (2D) LiDAR point data, a camera (40) mounted on the mobility (60) and configured to capture an image from the mobility (60), and a control unit (50) configured to - load a map to which 2D feature point data is bound, - to receive the 2D LiDAR point data from the LiDAR (10), - to receive the image from the camera (40), - to obtain three-dimensional (3D) observation feature point data from the image, - reduce one dimension of the 3D observation feature point data to 2D observation feature point data, - to bind the 2D observation feature point data to the 2D LiDAR point data and - to acquire an absolute position of the mobility based on the 2D observation feature point data linked to the 2D LiDAR point data and based on the map. [10] The system of claim 9, wherein the control unit (50) is further configured to - to select a calculated position of mobility (60), - compare the bound 2D observation feature point data with the 2D feature point data bound to the calculated map position, and - to select a final position of mobility (60) on the basis of the comparison. [11] The system of claim 10, wherein the control unit (50) is further configured to select the final position of the mobility (60) by checking a similarity of the bound 2D observation feature point data and the 2D feature point data bound to the calculated position of the map. [12] The system of claim 10 or 11, further comprising a motion data sensor attached to the mobility (60) and configured to capture motion data of the mobility (60), wherein the controller (50) is further configured to select the calculated position of the mobility (60) by locating the mobility (60) based on one or both of the 2D LiDAR point data and the motion data of the mobility (60). [13] The system of any one of claims 9 to 12, wherein the controller (50) is further configured to update the map to which the 2D feature point data is bound based on the absolute position of the mobility (60) and based on the 2D observation feature point data bound to the 2D LiDAR point data. [14] The system of claim 13, wherein the control unit (50) is further configured to store the updated map in a memory (52). [15] The system of any one of claims 9 to 14, wherein the control unit (50) is further configured to, prior to reducing the dimension of the 3D observation feature point data to the 2D observation feature point data, delete collision-free 3D observation feature point data that do not collide with the mobility (60) from the 3D observation feature point data. [16] The system of claim 15, wherein specifications of the mobility (60) are stored in a memory (52) and the control unit (50) is further configured to select the collision-free 3D observation feature point data based on the specifications of the mobility (60). [17] Mobility (60), which has: a drive motor, a wheel, a value sensor (20) configured to detect information about a rotation of the drive motor or the wheel, an inertial sensor (30) configured to detect information about a movement situation of the mobility (60), a light detection and ranging device (LiDAR) (10), a camera (40), a control unit (50), and a memory (52) storing instructions which, when executed by the control unit (50), cause the control unit (50) to: Loading a map to which two-dimensional (2D) feature point data is bound, Receiving 2D LiDAR point data from the LiDAR (10), Receiving an image from the camera (40), Obtaining three-dimensional (3D) observation feature point data from the image, Reducing one dimension of the 3D observation feature point data to 2D observation feature point data, Binding the 2D observation feature point data to the 2D LiDAR point data, Selecting a calculated position of the mobility (60) based on the 2D LiDAR point data, information about a rotation of the drive motor or the wheel detected by the value sensor (20), and information about a movement situation of the mobility (60) detected by the inertial sensor (30), and Acquiring an absolute position of the mobility (60) based on the 2D observation feature point data bound to the 2D LiDAR point data and based on the map.

Images