Methods for creating environmental maps for use in autonomous navigation of mobile robots

Abstract

In a method of creating an environment map for use in autonomous navigation of a mobile robot (10) in a map-based positioning system in a logistics environment, mobile objects (20, 200) of a goods management system having known positions are included into the environment map creation process.

Description

Technical Field

[0001] This invention relates to a method for creating an environmental map for use in autonomous navigation of a mobile robot in a map-based positioning system within a logistics environment. The invention also relates to a computer program for performing such a method, a machine-readable storage medium, and an electronic control device for performing the method, wherein such a computer program is stored on the storage medium. Background Technology

[0002] Currently, it has become standard practice in logistics centers for automated guided vehicles to transport goods in an automated manner. The navigation of these vehicles / robots takes place in structured environments, for example, based on markings on the ground.

[0003] Therefore, German publication 10 2014 100 658 A1 describes, for example, a tower crane that can be used in port facilities. This tower crane allows for the manipulation of containers, and includes a horizontal crossbeam with a crane trolley movable in the direction of travel, wherein the crane trolley is movable both along and laterally in the direction of travel and has a liftable load-bearing device for the container.

[0004] As the foundation for robot navigation, a variant of the so-called SLAM (Simultaneous Localization and Mapping) algorithm is typically used. In this case, the robot creates a map of its environment and simultaneously localizes itself within that map. Algorithms are used for this process to determine the robot's position relative to its environment. For this purpose, an environmental representation, i.e., a map, can be created using the onboard sensors of the mobile robot. This map will include the physical characteristics of the environment detected by the sensors, such as a laser scanner (LIDAR) or a camera. The localization map can, depending on the sensors and methods used, consist of a combination of originally measured points, salient points (features), or semantic objects (high-level features). In addition to simultaneous localization and map creation, mapping and localization may also be performed in separate steps, if necessary, using different vehicles or mobile robots. Summary of the Invention

[0005] Advantages of the present invention

[0006] This invention provides a method for creating an environment map for use in autonomous navigation of mobile robots in a map-based positioning system within a logistics environment, wherein mobile objects with known locations in a merchandise management system are included in the environment map creation process. This method can improve the robustness and positioning accuracy of map-based positioning systems, for example, in port settings, thereby indirectly improving the navigation quality of the corresponding robots, particularly transport robots, such as drivable cranes or other vehicles. Mobile robots can also be understood as flyable and / or autonomously operating robots, such as drones, particularly multi-rotor helicopters. The flyable robots can also be configured as air taxis for transporting people or passengers.

[0007] The proposed method thus provides a clear improvement in localization and navigation in variable environments. Typically, navigation of autonomous transport robots in logistics environments, such as port facilities, is difficult because containers are frequently moved within these facilities, making it challenging to create persistent maps based on environmental representations. The proposed method addresses this problem by incorporating known mobile objects—whose locations are stored in a merchandise management system—into the environmental map creation process. Here, existing information from the merchandise management system regarding the locations of mobile objects, such as containers, pallets, or other prominent objects, is used and included in the environmental map creation process within the scope of the localization system. Using this information from the merchandise management system, it is possible to, for example, identify whether a specific object is a mobile object. Furthermore, the changing locations of mobile objects can be directly considered in the environmental map.

[0008] A merchandise management system is, in this context, a system that reflects the flow of goods within a company's business processes. For the purposes of this invention, the merchandise logistics of the merchandise management system is crucial, as it is used to organize warehouse management. This warehouse management system includes information about the location of objects to be retrieved from or stored in the warehouse. Therefore, it is possible to retrieve information about the location of, for example, containers, pallets, or other mobile objects within the merchandise management system. Thus, for example, port management often utilizes a merchandise management system, in which the location of containers being transferred within port facilities is monitored in real time.

[0009] The information is incorporated into the creation of an environmental map used for autonomous navigation of a mobile robot, according to the method of the present invention.

[0010] This method is therefore particularly advantageous for use with Automated Guided Vehicles (AGVs), such as those transporting containers in ports. Furthermore, the method is also applicable to other automated vehicles or other mobile robots used in logistics environments.

[0011] Preferably, the logistics environment is a port facility. Port facilities are typically very dynamic environments because large numbers of containers are usually moved, rearranged, or transshipped within the port. In this case, precision-working cranes are used, for example, to pick up, transport, and place containers or other units into predetermined locations. Container handling is usually controlled via a merchandise management system. Such cranes or other mobile transport robots, for example, operate based on laser positioning and are particularly advantageously suited to applying the proposed methods.

[0012] In principle, it is possible that the environmental map used for navigation of a mobile robot is created solely based on location information about mobile objects from a merchandise management system. This map can be continuously compared with information from the merchandise management system, allowing for timely or real-time updates on changes in the location of objects, such as containers, within the environmental map. The mobile robot can then locate itself and navigate precisely within this dynamic map.

[0013] In a particularly preferred configuration of the method, mapping of mobile objects can be additionally performed based on sensor data from the mobile robot, wherein these objects detected by the mobile robot are compared with known location data from a merchandise management system to create an accurate environmental map. Preferably, the process is performed within the scope of simultaneous localization and mapping of the mobile robot, particularly within the scope of SLAM algorithms.

[0014] In an advantageous configuration of the proposed method, a dynamic semantic map is generated based on the known locations of mobile objects within the merchandise management system when creating the environmental map. In this case, it is not mandatory to map the mobile objects using the mobile robot's sensing mechanisms. Instead, such a dynamic semantic map can be generated solely based on information from the merchandise management system. Particularly advantageous in this case is that the known and changing locations of containers or other mobile objects from the merchandise management system can be reflected in the dynamic semantic map in a timely or real-time manner, ensuring that a current environmental map is always available for precise localization and navigation of the mobile robot.

[0015] In another embodiment of the method, it may be alternatively or additionally specified that, when creating the environment map, not only static objects of the merchandise management system are mapped, but also mobile objects of the merchandise management system are mapped. In this configuration, additional layers can be extended, for example, to a conventionally generated location map, where these additional layers represent dynamic objects of the merchandise management system or their locations, and where the conventionally generated location map consists, for example, of a static map based on point clouds. Mobile robots or other automated vehicles can therefore utilize a combination of static and dynamic environment maps for their own localization and navigation. In this case, the resulting environment map has, to some extent, two layers. The first layer represents the static environment map. The other layers represent mobile objects of the merchandise management system, and these other layers are dynamically adaptable.

[0016] In another preferred embodiment of the proposed method, the conventionally generated location map, which is typically based on static objects in the environment, can be bereinigenized such that mobile objects of the merchandise management system are filtered out during mapping when creating the environment map. This is based on the fact that the sensing mechanisms of mobile robots are generally unable to distinguish between static and mobile objects. When mapping by mobile robots or mapping vehicles, conventional methods typically cannot prevent the detection of mobile objects and their recording in the static map. This leads to problems for the navigation of mobile robots when the positions of the corresponding mobile objects change. The proposed method solves this problem by including information about the known positions of mobile objects in the merchandise management system, thus filtering out such dynamic objects during the mapping process and preventing them from being recorded in the environment map. In this embodiment of the method, the environment map created according to the proposed method partially omits the mobile objects in order to create the environment map, so that the navigation of the mobile robot is based solely on the unchanging positions of the static objects in the environment.

[0017] To create an environment map, it is advantageous to project the position of the moving object onto a reference coordinate system. Therefore, it is possible to convert moving objects with dynamic positions into a map-like representation, which can be used to locate the mobile robot using conventional methods, such as scan-matching.

[0018] The invention also includes a computer program configured to perform the steps of the proposed method. This computer program can be used, particularly in the context of map-based positioning systems, in logistics environments, especially in port facilities. Such a computer program can be configured, for example, as an add-on module to a merchandise management system to further improve the autonomous navigation of mobile robots, such as autonomous transport vehicles or mobile cranes. The invention also includes: a machine-readable storage medium on which such a computer program is stored; and an electronic control device configured to perform the steps of the proposed method.

[0019] The advantages mentioned earlier also apply, in a corresponding manner, particularly to mobile robots, for creating maps for use in autonomous navigation. Mobile robots can also be configured as flying and / or autonomously operating robots, such as drones, especially multi-rotor helicopters. Flying robots can also be constructed as air taxis for transporting people or passengers.

[0020] Other features and advantages of the invention will become apparent from the following description of embodiments taken in conjunction with the accompanying drawings. In this case, each feature is implemented individually or in combination with each other. Attached Figure Description

[0021] The accompanying diagram illustrates a current laser scan of a mobile robot autonomously navigating within the dynamic environment of a port facility. Detailed Implementation

[0022] This figure illustrates the problem of localization for a mobile robot 10 autonomously navigating within a port facility. The port facility is characterized by a changing environment, caused by the continuous repositioning of containers 20 and 200. The scenario schematically shown in the figure is based on a conventional method, according to which the mobile robot 10 is located in the port environment using a static map, such as a point cloud (points in a raw LiDAR map), represented by dotted lines. The mobile robot 10 is equipped, for example, with a laser scanner and generates a laser scan (currently measured LiDAR points), represented here by dashed outlines. However, for localization, other sensors, such as cameras, can also be used on the mobile robot 10, alternatively or additionally. This localization is performed by means of so-called scan matching, and navigation is based on this localization. The current position and orientation of the mobile robot 10 are estimated such that the current laser scan (dashed outline) matches the point cloud (dotted outline) of the static map as well as possible. In this case, the static map can, for example, consist of point features (e.g., SIFT features, Scale Invariant Feature Transform), which are rediscovered as corresponding locations in the current camera image. The estimated camera position and orientation are then estimated such that the backprojection error at all corresponding locations is minimized. Thus, the difference between the expected location and the actually identified location of the point features in the image is kept as small as possible.

[0023] Some containers 20 and 200 are identifiable in the current laser scan (dashed lines) and static map (dots), but these containers are located in different positions in the static, and therefore outdated, map compared to the current laser scan. Container 20 is one such container that was present in the original (static) map but is no longer there. Container 200 is one such container that is currently present but not present in the original (static) map. This leads to problems in the localization and navigation of the mobile robot 10. These problems are solved by the proposed method, which compares the objects detected by the laser scan of the mobile robot 10 with the merchandise management system used in the logistics environment. Here, the merchandise management system is thus connected to a localization or mapping system. Therefore, the environmental map can be updated so that changes in the positions of containers 20, 200, or other mobile objects whose own positions are known are traced and corrected. The environmental map is therefore always current in principle and does not present any problems in the localization and navigation of the mobile robot 10.

[0024] For example, it can be stipulated that mobile objects 20, 200, i.e., containers in a merchandise management system, are projected onto a reference coordinate system and thus transformed into a map-like representation. This map-like representation can be created, for example, solely based on information from the merchandise management system. This dynamic semantic map can be used to locate and navigate the mobile robot 10 in a manner known per se, for example by means of scan matching and evaluation of point features, wherein the dynamic semantic map contains the current position of the objects in the merchandise management system.

[0025] Such a map can also be combined with a traditional static location map, wherein static objects 30 of the environment are used to create the map. Depending on the method, this can be done by accumulating not only points or point features from the static map but also points or point features from the dynamic map. In particular, the positions of mobile objects 20, 200, i.e., containers, can be considered such that their positions in the static map are corrected by overlaying the current position 200 with the outdated position 20.

[0026] Furthermore, considering the application of the proposed method to the dynamic or mobile objects 20, 200 in the conventional location map clearing product management system, in order to filter out mobile objects 20, 200 during data processing for mapping, all measurements, i.e., points or other visual point features such as laser scans, are examined to determine whether the measurements can be assigned to mobile objects. If the detected point or point feature is a mobile object 20, 200, the detected point or point feature can be discarded or filtered out for the creation of the static location map. Therefore, for example, the static map of containers 20, 200 can be cleared, thereby using only static objects 30 for the localization and navigation of the mobile robot 10.

Claims

1. A method for creating an environmental map for use in autonomous navigation of a mobile robot (10) in a map-based positioning system within a logistics environment, characterized in that, Existing information from the commodity management system regarding the location of the mobile objects (20, 200) with known locations in the commodity management system is used and included in the environmental map creation process within the scope of the positioning system. The environmental map is continuously compared with information from the commodity management system, so that the changing locations of the mobile objects in the environmental map are taken into account in a timely or real-time manner. The mobile objects are containers or other units in the logistics environment that can be rearranged or transferred. A dynamic semantic map is generated based on the known locations of the mobile objects (20, 200) in the commodity management system when creating the environmental map. The resulting environmental map has two layers to a certain extent: a first layer represents a static environmental map, and other layers represent the mobile objects in the commodity management system. These other layers are dynamically adaptable. The mobile objects (20, 200) in the commodity management system are filtered out during mapping when creating the environmental map.

2. The method according to claim 1, characterized in that, The logistics environment refers to port facilities.

3. The method according to any one of the preceding claims, characterized in that, The creation of the environmental map is based on sensor data from the mobile robot (10).

4. The method according to any one of the preceding claims, characterized in that, The creation of the environmental map is carried out within the scope of the synchronous localization and mapping of the mobile robot (10).

5. The method according to any one of the preceding claims, characterized in that, When creating the environment map, the mapping of the static object (30) is supplemented by the mapping of the mobile objects (20, 200) of the commodity management system.

6. The method according to any one of the preceding claims, characterized in that, The positions of the moving objects (20, 200) are projected onto the reference coordinate system.

7. A computer program product comprising a computer program that, when executed on a processor, performs the steps of the method according to any one of claims 1 to 6.

8. A machine-readable storage medium on which the computer program according to claim 7 is stored.

9. An electronic control device configured to perform the steps of the method according to any one of claims 1 to 6.

10. A mobile robot (10) having the electronic control device according to claim 9.

Images