Closed-loop detection method, device, robot and storage medium

Abstract

The embodiments of the present application disclose a closed-loop detection method, device, robot and storage medium, which relate to the field of robot navigation technology. The method comprises: identifying a current tag of the robot at a current position; generating target feature data according to the position information of the current tag; performing closed-loop detection on the movement path of the robot according to the target feature data; through the above technical scheme, the efficiency and accuracy of closed-loop detection are improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of robot navigation technology, and in particular to a closed-loop detection method, device, robot and storage medium. Background Art

[0002] Simultaneous Localization and Mapping (SLAM) is a basic problem and research hotspot in the field of robot navigation. In the SLAM process, the robot achieves self-localization and builds an environmental map at the same time. Due to the lack of prior knowledge and environmental uncertainty, the robot needs to determine whether the current position is in an environmental area that has been visited during walking, and use this as a basis for determining whether the environment needs to be updated, which is the closed-loop detection problem.

[0003] In recent years, more and more closed-loop detection algorithms based on the bag-of-words model have emerged. This technology uses an offline training process to extract a large number of feature descriptors from training images, cluster them, and generate a visual dictionary based on the cluster centers. As the image stream is continuously input during the movement of the robot, a database is generated online using the visual dictionary. Then, the similarity of image features between the current key frame and the candidate matching frames in the database is calculated to determine whether a closed loop should occur.

[0004] However, the above method takes a lot of time to extract image features, and as the robot's moving range expands, the amount of calculation will gradually increase, eventually leading to a sharp decrease in the real-time and accuracy of closed-loop detection. Summary of the invention

[0005] The present application provides a closed-loop detection method, device, robot and storage medium to improve the efficiency and accuracy of closed-loop detection.

[0006] In a first aspect, an embodiment of the present application provides a closed-loop detection method, the method comprising:

[0007] Identify the current tag of the robot at its current location;

[0008] Generate target feature data according to the position information of the current tag;

[0009] A closed-loop detection is performed on the movement path of the robot according to the target feature data.

[0010] In a second aspect, an embodiment of the present application further provides a closed-loop detection device, the device comprising:

[0011] The tag recognition module is used to recognize the current tag of the robot at the current position;

[0012] A feature data generating module, used to generate target feature data according to the position information of the current tag;

[0013] The closed-loop detection module is used to perform closed-loop detection on the movement path of the robot according to the target feature data.

[0014] In a third aspect, an embodiment of the present application further provides a robot, the device comprising:

[0015] one or more processors;

[0016] a storage device for storing one or more programs,

[0017] When the one or more programs are executed by the one or more processors, the one or more processors implement any one of the closed-loop detection methods provided in the first aspect embodiment.

[0018] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any one of the closed-loop detection methods provided in the embodiment of the first aspect.

[0019] The embodiment of the present application identifies the current tag of the robot at the current position, generates target feature data according to the position information of the current tag, and performs closed-loop detection on the movement path of the robot according to the target feature data; through the above-mentioned technical scheme, based on at least one identifiable tag pre-set in the robot's working environment, the target feature data is calculated by identifying the label in the scene, and based on the target feature data, it is detected whether there is a closed loop in the movement path of the robot, thereby solving the complexity problem of the image feature matching method; at the same time, with the help of the uniqueness of the label itself, it is possible to exclude the situation where the scene features are similar but are actually different scenes, thereby effectively improving the efficiency and accuracy of closed-loop detection and meeting the requirements of closed-loop detection for real-time and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a flow chart of a closed-loop detection method provided in Example 1 of the present application;

[0021] Figure 2 is a flow chart of a closed-loop detection method provided in Embodiment 2 of the present application;

[0022] Figure 3 This is a schematic diagram of a robot moving path for closed-loop detection provided in Example 2 of the present application;

[0023] Figure 4 is a flow chart of a closed-loop detection method provided in Example 3 of the present application;

[0024] Figure 5 is a schematic diagram of a closed-loop detection device provided in Embodiment 4 of the present application;

[0025] Figure 6 This is a schematic diagram of a robot provided in Example 5 of the present application. DETAILED DESCRIPTION

[0026] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for ease of description, only the parts related to the present application, rather than all structures, are shown in the accompanying drawings.

[0027] It should be mentioned before discussing the exemplary embodiments in more detail that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the steps as sequential processes, many of the steps therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the steps can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the accompanying drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0028] Embodiment 1

[0029] Figure 1 A flowchart of a closed-loop detection method provided in Example 1 of the present application. The present application embodiment can be applied to the case where a robot detects whether a moving path has a closed loop during movement. The method can be performed by a closed-loop detection device, which can be implemented by software and / or hardware and specifically configured in the robot.

[0030] See also Figure 1 , the closed-loop detection method provided in the embodiment of the present application includes:

[0031] S110, identifying the current tag of the robot at the current position.

[0032] Among them, robots refer to intelligent robots that can autonomously or interactively perform various anthropomorphic tasks in various working environments. Specific types of robots usually include but are not limited to food delivery robots, inspection robots, and advertising robots.

[0033] A tag is a unique identifiable tag pre-set in the robot's operating environment. The specific type of the tag can be a dot matrix tag, a QR code tag, or an RFID (Radio Frequency Identification) tag.

[0034] In this embodiment, at least one tag may be pre-set in the robot's operating environment, and of course, multiple tags may be pre-set according to the operating scenario. Typically, the tag may be affixed above a building, such as on an indoor ceiling, to prevent human tampering and damage.

[0035] Specifically, when the robot moves in a scene with preset labels, it can use sensors such as cameras carried by the robot itself to collect images of the surrounding environment and identify labels from the collected images, or use specific label recognition modules such as radio frequency identification modules to identify current labels in the current environment, thereby obtaining the current label of the robot at its current position.

[0036] S120: Generate target feature data according to the position information of the current tag.

[0037] The location information of the tag refers to the location information of the tag in the scene. For example, the location information of the current tag can be represented by three-dimensional coordinates L1 (x1, y1, z1), where L1 represents the current tag.

[0038] In this embodiment, the location information of the tag can be determined by the robot through its own positioning system. For example, the location information of the tag can be: the robot obtains the location information of the tag through the odometer and the camera sensor, wherein the robot obtains the location information of the robot according to its own odometer, obtains the relative position of the robot and the tag after collecting the tag through the robot camera, and calculates the location information of the tag according to the robot coordinate information and the relative position of the robot and the tag.

[0039] The target feature data refers to the feature data associated with the current tag, generated according to the position information of the current tag, and capable of characterizing the current environment or scene of the robot.

[0040] In this embodiment, unlike the prior art method of extracting features from the entire image, the present application generates target feature data based on the position information of the current tag.

[0041] It is understandable that since it is only necessary to generate feature data based on the labels in the scene without processing the entire image, the amount of computation is very small and real-time calculation can be achieved, so that closed-loop detection of the robot's movement path can be performed based on the target feature data.

[0042] In this embodiment, for simplicity, the target feature data can be generated only based on the position information of the current tag; of course, in order to ensure the accuracy of the target feature data, other information can be referenced on the basis of the current tag, and the target feature data can be generated together with the position information and other information of the current tag. The other information can be other feature data extracted from the current scene image of the robot, or label information appearing in the historical scene, etc.

[0043] It is understandable that this embodiment does not impose any restrictions on the process of generating target feature data, and it can be selected according to actual conditions.

[0044] S130: Perform closed-loop detection on the robot's moving path according to the target feature data and the label identification.

[0045] Among them, closed-loop detection, also known as loop detection, means that the robot returns to the starting point after traveling a circle based on sensor information to determine whether the robot’s current position has been described accordingly in the created map.

[0046] Usually, as the robot's moving path continues to extend, there will be some cumulative errors in the robot's mapping process. In addition to using local optimization and global optimization to adjust, closed-loop detection can also be used for optimization.

[0047] In this embodiment, the tag identification of the current tag can be determined according to the dot matrix tag, the two-dimensional code tag or the RFID tag, and the tag identification is unique.

[0048] Specifically, when performing closed-loop detection on the robot's moving path, the target feature data can be first matched with the scene feature data stored in the database. If the matching results are consistent, the corresponding matching scene label is output; further, it is determined whether the label identifiers are consistent based on the label identifiers, that is, the current label identifier in this embodiment is matched with the matching scene label identifier to exclude the situation where the scene features are similar but are actually different scenes.

[0049] It is understandable that if the closed-loop detection is successful, useful closed-loop information can be derived based on the closed-loop detection results, such as obtaining a map with consistent information, to reduce the cumulative error of the sensor, which can help the robot avoid obstacles and navigate more accurately and quickly; if the closed-loop detection fails or is unsuccessful, erroneous closed-loop information may be obtained or used, which will not only interfere with subsequent map optimization processing, but may even completely destroy the existing map creation results.

[0050] The embodiment of the present application identifies the current tag of the robot at the current position, generates target feature data according to the position information of the current tag, and performs closed-loop detection on the movement path of the robot according to the target feature data; through the above-mentioned technical scheme, based on at least one identifiable tag pre-set in the robot's working environment, the target feature data is calculated by identifying the label in the scene, and based on the target feature data, it is detected whether there is a closed loop in the movement path of the robot, thereby solving the complexity problem of the image feature matching method; at the same time, with the help of the uniqueness of the label itself, it is possible to exclude the situation where the scene features are similar but are actually different scenes, thereby effectively improving the efficiency and accuracy of closed-loop detection and meeting the requirements of closed-loop detection for real-time and accuracy.

[0051] Embodiment 2

[0052] Figure 2 A flow chart of a closed-loop detection method provided in Embodiment 2 of the present application. This embodiment is an optimization of the above-mentioned solution based on the above-mentioned embodiment.

[0053] Furthermore, the operation of "generating target feature data according to the position information of the current tag" is refined into "generating the target feature data according to the position information of the current tag and the position information of at least one historical tag" to clarify the generation process of the target feature data.

[0054] The explanations of the terms that are the same as or corresponding to the above embodiments will not be repeated here.

[0055] See also Figure 2 , the closed-loop detection method provided in this embodiment includes:

[0056] S210, identifying the current tag of the robot at the current position.

[0057] S220: Generate target feature data according to the location information of the current tag and the location information of at least one historical tag.

[0058] Among them, the historical tag refers to the tag that appeared in the historical scene during the movement of the robot, that is, the robot recognizes at least one historical tag on the robot's movement path and before passing the position of the current tag.

[0059] In this embodiment, the number of tags that form the generated target feature data can be flexibly configured. Generally, the error of feature matching with a single tag may be large. For example, if the tag is recognized incorrectly, a mismatch may occur. Therefore, in this embodiment, the target feature data is generated based on the current tag and at least one historical tag, that is, at least two tags, so that the probability of mismatch is reduced exponentially, because the probability of multiple tags being recognized incorrectly is low.

[0060] In some embodiments, the location information of the current tag and the location information of at least one historical tag may be directly used as target feature data.

[0061] Or optionally, in order to further ensure the accuracy of the target feature data, the target feature data is generated according to the position information of the current tag and the position information of at least one historical tag, including: generating at least one set of tag pairs according to the current tag and the at least one historical tag, wherein the current tag and the historical tag are different; generating relative position information according to the position information of each tag in the tag pair, wherein the position information of the tag can be the position information of the tag calculated by the robot according to the positioning, and the position information includes coordinate information and angle information, wherein the coordinate information refers to the coordinate information in the robot coordinate system, and the angle information refers to the inclination angle information of the tag and the coordinate plane. The coordinate information is used as the position feature point of the tag, and the relative position information can be the relative geometric relationship between the position feature points of the two tags, such as direction or distance; the target feature data is generated according to at least one of the relative position information.

[0062] The relative position information refers to the relative position information between the position information of each tag in the tag pair. Figure 3 The schematic diagram of a robot moving path for closed-loop detection is shown, and the relative position information between each tag is generated, that is, the connecting line with arrows in the figure. Among them, the figure exemplarily shows the connecting line of the relative position information of tag L1 and tag L2, the connecting line of the relative position information of tag L2 and tag L3, and the connecting line of the relative position information of tag L3 and tag L4.

[0063] For example, assuming there are three historical tags L1, L2 and L3, the position information of the current tag is represented by tag L4 (x4, y4, z4), and the position information of the historical tags is represented by tags L1 (x1, y1, z1), L2 (x2, y2, z2) and tag L3 (x3, y3, z3). Then the relative position information between the tags in the tag pair may include: the relative position information between tag L1 and tag L2, the relative position information between tag L2 and tag L3, and the relative position information between tag L3 and tag L4.

[0064] It is understandable that if there is a measurement error in the sensor, the detection of the tag position information will be incorrect, but the relative position information between the tags usually remains unchanged. The relative position information between the tags in the tag group can be used to determine the target feature data, making the generated target feature data more accurate.

[0065] Optionally, the relative position information includes relative distance and / or relative direction.

[0066] Here, the relative distance refers to the distance between two labels.

[0067] The relative direction refers to the direction of one label relative to the other label between two labels, such as the direction of the latter label relative to the former label.

[0068] It can be understood that the relative position information between each tag can be measured by a variety of indicators. For the sake of simplicity, the relative position information between each tag can be determined based on the relative distance or relative direction between each tag. If the position relationship between the tags is more complex, at the same time, in order to improve the accuracy of the characterization of the relative position relationship, the relative position information between each tag can be determined by both relative distance and relative direction.

[0069] In some embodiments, on the basis of expressing the relative position information by relative distance and / or relative direction, the position information of the tag itself can also be taken into account, such as taking the position information of one of the tags into account, or taking the position information of both tags into account at the same time, thereby further improving the accuracy of the relative position information between the tags.

[0070] In this embodiment, for the process of generating target feature data, at least one set of label pairs can be generated from the current label and at least one historical label, and relative position information can be generated based on the position information of each label in each label pair. At this time, multiple relative position information can be obtained. The present application can select at least one relative position information from the above-mentioned multiple relative position information as the target feature data. Of course, all relative position information can also be used as target feature data.

[0071] Continuing with the above example, tag L1, tag L2 and tag L3 can form a tag pair of tag L1 and tag L2, a tag pair of tag L1 and tag L3, and a tag pair of tag L2 and tag L3; accordingly, the relative position information of tag L1 and tag L2, the relative position information of tag L1 and tag L3, and the relative position information of tag L2 and tag L3 can be generated.

[0072] Optionally, generating at least one set of tag pairs according to the current tag and the at least one historical tag includes: identifying tags with adjacent positions; and selecting two tags with adjacent positions in sequence to generate the tag pairs.

[0073] Among them, adjacent recognition positions refer to the order in which the robot observes the positions of the tags.

[0074] Continuing with the above example, if label L3 is the current label, label L2 is the label observed in the previous scene of label L3, and label L2 is the label observed in the previous scene of label L1, then the corresponding label pairs that can be generated include: a label pair consisting of label L1 and label L2, and a label pair consisting of label L2 and label L3.

[0075] It is understandable that the relative position information generated by adjacent tag pairs is more accurate. If two tags separated by multiple tags are used as a tag pair to generate relative position information, there will be a large error because the sensor's detection of the tag position will have cumulative errors.

[0076] Typically, the target feature data is generated based on the location information of the current tag and the location information of at least one historical tag, including: identifying tags with adjacent locations to sequentially form a target tag connection line of a connection line relationship, and generating the target feature data based on the connection line attributes of the target tag connection line; wherein the connection line attributes include location information, direction information, and tag ID information at both ends. The tag ID information includes unique identification information such as the region and location information of the tag. Through the tag ID information, the situation where different ID pairs appear but the connection lines are familiar with the same can be avoided.

[0077] The connection line attributes include not only the position information and the direction information, but also the length information.

[0078] It can be understood that the target feature data is specifically refined here, and the target feature data is composed of the connection line attributes of two adjacent tags.

[0079] In some embodiments, there may be situations where historical tag detection fails or is blocked. At this time, target feature data can still be generated based on the location information of the current tag and the location information of at least one historical tag. This does not affect the generation of the target feature data and has strong compatibility. The only difference is that the target feature data is no longer generated based on pairwise adjacent tags in the actual scene.

[0080] For example, taking the target feature data composed of 4 labels as an example, if the labels observed by the robot in the first circle are (L1, L2, L3, L4, L5) in sequence, then a total of 5 groups of feature data with the number of combinations C (5, 4) can be generated. In the second circle, due to the occlusion of the label point L3, the labels that the robot can observe are (L1, L2, L4, L5). At this time, the target feature data can still be generated based on (L1, L2, L4, L5) to perform closed-loop detection on the robot's moving path, because (L1, L2, L3, L4, L5) contains the combination of (L1, L2, L4, L5).

[0081] S230: Perform closed-loop detection on the robot's moving path according to the target feature data.

[0082] Based on the above embodiments, the embodiment of the present application clarifies the generation process of target feature data, and generates target feature data according to the position information of the current tag and the position information of at least one historical tag; through the above technical scheme, compared with the process of generating target feature data only according to the current tag, at least two tags consisting of the current tag and at least one historical tag can make the generated target feature data more accurate, so that according to the above-determined target feature data, the movement path of the robot is closed-loop detected, the result of the closed-loop detection can be made more accurate, and the accuracy of the closed-loop detection is further improved.

[0083] Embodiment 3

[0084] Figure 4 This is a flow chart of a closed-loop detection method provided in Example 3 of the present application. This embodiment is an optimization of the above-mentioned solution based on the above-mentioned embodiment.

[0085] Furthermore, the operation of "performing closed-loop detection on the movement path of the robot according to the target feature data" is refined into "determining standard feature data corresponding to the target feature data in a preset matching library; and performing closed-loop detection on the movement path of the robot according to the matching situation between the target feature data and the standard feature data" to improve the closed-loop detection process.

[0086] The explanations of the terms that are the same as or corresponding to the above embodiments will not be repeated here.

[0087] See also Figure 4 , the closed-loop detection method provided in this embodiment includes:

[0088] S310, identifying the current tag of the robot at the current position.

[0089] S320: Generate target feature data according to the location information of the current tag and the location information of at least one historical tag.

[0090] S330: Determine standard feature data corresponding to the target feature data in a preset matching library.

[0091] The scene correspondence between the standard feature data and the target feature data refers to feature data that can characterize the scene or environment that the robot has previously visited. The preset matching library refers to a feature database that stores all scenes that the robot has previously visited.

[0092] Specifically, during the movement of the robot, the robot can record and save the scenes with labels through the sensors carried by itself. For example, the robot can save the multiple labels (L1, L2, L3, L4, L5) observed in the first round in the order of collection, and save the scene feature data represented by the labels according to the preset feature data generation rules, and store them in the preset matching library. Among them, the preset feature data generation rules can be to generate feature data according to the relative position information between each label. Typically, feature data can be generated according to the relative position information of adjacent labels.

[0093] Optionally, determining the standard feature data corresponding to the target feature data in the preset matching library includes: generating a standard label group according to the identification information of each label corresponding to the target feature data; and generating the standard feature data according to the standard position information of each label in the standard label group during the movement of the robot's moving path.

[0094] The identification information of the tag may be a number pre-set for the tag, such as the number of tag L1 may be represented by L1, which represents the tag itself and is unique. Usually, the identification information of the tag is associated with the scene or environment where the robot is located.

[0095] In this embodiment, if the labels determined according to the identification information of each label corresponding to the target feature data include labels L1, L2 and L3, then the labels L1, L2 and L3 originally collected by the robot can be formed into a standard label group, and standard feature data can be generated according to the standard position information of each label in the standard label group.

[0096] It is understandable that the generation of standard feature data needs to correspond to the target feature data, that is, the current scene of the robot needs to be feature matched with the scenes it has visited before, in order to determine whether the robot's movement path is closed.

[0097] Optionally, the preset matching library is constructed in the following manner: constructing at least one standard tag group based on the identification information of each standard tag; generating the standard feature data based on the position information of each standard tag in the standard tag group; and constructing the preset matching library based on at least one group of the standard feature data.

[0098] For example, there are 5 standard labels observed by the robot in the first circle, which are represented by the label set (L1, L2, L3, L4, L5). If the generation of feature data requires 4 labels, then a total of 5 standard label groups with the number of combinations C (5, 4) can be constructed, including (L1, L2, L3, L4), (L1, L2, L4, L5), (L1, L3, L4, L5), (L1, L2, L4, L5) and (L2, L3, L4, L5). Accordingly, 5 sets of standard feature data can be generated.

[0099] In this embodiment, all the generated standard feature data can be stored in the preset matching library. Of course, part of the generated standard feature data can also be stored in the preset matching library, such as storing the standard feature data generated by the standard tag groups (L1, L2, L3, L4) and (L2, L3, L4, L5) in the preset matching library. The embodiment of the present application does not impose any restrictions on the selection and storage of standard feature data, and can be set according to actual conditions, such as considering factors such as storage space size and matching efficiency.

[0100] It is understandable that as the robot's moving range expands, more and more labels will be observed. The robot can generate rules based on preset feature data, continuously expand and save the scene feature data represented by the labels, and continuously improve the preset matching library.

[0101] S340: Perform closed-loop detection on the movement path of the robot according to the matching of the target feature data and the standard feature data, and according to the matching of the tag identifier corresponding to the target feature data and the tag identifier corresponding to the standard feature data.

[0102] Optionally, the target feature data is matched one by one with the corresponding relative position information in the standard feature data. If the matching errors are all smaller than a preset matching threshold, it can be determined that the matching result between the target feature data and the standard feature data is successful.

[0103] Among them, the preset matching threshold can be selected according to the actual situation. Of course, different preset matching thresholds can also be selected for the relative position information formed by different tags. The specific setting of the preset matching threshold can be determined according to the actual situation.

[0104] It is understandable that if the feature data contains multiple sets of relative position information, then when performing feature matching, it is necessary to match the target feature data with the corresponding relative position information in the standard feature data one by one, and determine whether the robot's movement path is closed based on the matching results.

[0105] In this embodiment, if the matching result between the target feature data and the standard feature data is a failure, the robot's moving path does not close in a loop; if the matching result between the target feature data and the standard feature data is a success, and the matching result between the tag identifier corresponding to the target feature data and the tag identifier corresponding to the standard feature data is a matching failure, the robot's moving path does not close in a loop; if the matching result between the target feature data and the standard feature data is a success, and the matching result between the tag identifier corresponding to the target feature data and the tag identifier corresponding to the standard feature data is a matching success (that is, the identifiers are consistent), the robot's moving path closes in a loop.

[0106] It is understandable that, with the uniqueness of the label itself, it is possible to exclude situations where the scene features are similar but they are actually different scenes.

[0107] Typically, the closed-loop detection of the robot's moving path according to the matching of the target feature data with the standard feature data, and according to the matching of the tag identifier corresponding to the target feature data with the tag identifier corresponding to the standard feature data, includes: sequentially forming target tag connecting lines with connecting line relationships with tags whose identification positions are adjacent, and generating target feature data based on the connecting line attributes of the connecting lines of the position feature points of the target tags; wherein the connecting line attributes include position information and direction information;

[0108] Accordingly, a closed-loop detection is performed on the moving path of the robot, including:

[0109] storing the second connection line attributes of the position feature points of the two target tags to generate standard feature data;

[0110] The target feature data is matched with the standard feature data, and the tag identifier corresponding to the target feature data is matched with the tag identifier corresponding to the standard feature data, and based on the matching result, a closed-loop detection is performed on the movement path of the robot, and the position information of the target tag is determined according to a predetermined rule.

[0111] Optionally, if the result of the closed-loop detection is that the matching result is within a predetermined error range and the actual position of the target tag remains unchanged, the position information of the two tags can be fused through a predetermined rule, such as the G2O (general graph optimization) algorithm in the SLAM field to finally determine the position information of the target tag.

[0112] Continue to see Figure 3 ,The figure takes the feature data composed of four tags (L1, L2, L3, L4) as an example, by matching the feature data formed by the four tags, closed-loop detection of the robot’s moving path can be achieved.

[0113] Specifically, see Figure 3 During the movement, the robot can obtain the position information of each tag in real time through the camera carried by the robot itself. When the robot passes through the tags L1-L4 for the first time, the corresponding feature data {L1-L2, L2-L3, L3-L4} can be generated, that is, the dotted line connecting edges in the figure, and the observed tags are represented by squares; when the robot loops back, it will see the four tags L1-L4 again in turn, and new feature data {L1*-L2*, L2*-L3*, L3*-L4*} will be formed, that is, the solid line connecting edges in the figure, and the observed tags are represented by triangles; if the relative position information of the above three pairs of connecting edges are all within the error range, that is, the relative position information of L1-L2 and L1*-L2*, L2-L3 and L2*-L3*, and L3-L4 and L3*-L4* are all within the preset error range, it can be identified as feature matching; if the matching result is a match, it can be determined that the robot's moving path is closed.

[0114] Based on the above-mentioned embodiments, the embodiments of the present application further refine the closed-loop detection process. First, the standard feature data corresponding to the target feature data in the preset matching library is determined, and then the robot's moving path is closed-loop detected based on the matching situation between the target feature data and the standard feature data. Through the above-mentioned technical scheme, the standard feature data corresponding to the target feature data is determined from the preset matching library, and the robot's moving path is closed-loop detected based on the matching situation between the target feature data and the standard feature data, thereby improving the closed-loop detection process and making it possible to fully detect whether the robot's moving path is closed-loop.

[0115] Embodiment 4

[0116] Figure 5 Schematic diagram of a closed-loop detection device provided in Embodiment 4 of the present application. Figure 5 , an embodiment of the present application provides a closed-loop detection device, which includes: a tag recognition module 410, a feature data generation module 420 and a closed-loop detection module 430.

[0117] The tag identification module 410 is used to identify the current tag of the robot at the current position;

[0118] The feature data generating module 420 is used to generate target feature data according to the position information of the current tag;

[0119] The closed-loop detection module 430 is used to perform closed-loop detection on the movement path of the robot according to the target feature data.

[0120] The embodiment of the present application identifies the current tag of the robot at the current position, generates target feature data according to the position information of the current tag, and performs closed-loop detection on the movement path of the robot according to the target feature data; through the above-mentioned technical scheme, based on at least one identifiable tag pre-set in the robot's working environment, the target feature data is calculated by identifying the label in the scene, and based on the target feature data, it is detected whether there is a closed loop in the movement path of the robot, thereby solving the complexity problem of the image feature matching method; at the same time, with the help of the uniqueness of the label itself, it is possible to exclude the situation where the scene features are similar but are actually different scenes, thereby effectively improving the efficiency and accuracy of closed-loop detection and meeting the requirements of closed-loop detection for real-time and accuracy.

[0121] Furthermore, the feature data generating module 420 includes:

[0122] The feature data generating submodule is used to generate the target feature data according to the location information of the current tag and the location information of at least one historical tag.

[0123] Furthermore, the feature data generation submodule includes:

[0124] a label pair generating unit, configured to generate at least one set of label pairs according to the current label and the at least one historical label;

[0125] A position information generating unit, used for generating relative position information according to the position information of each tag in the tag pair;

[0126] The feature data generating unit is used to generate the target feature data according to at least one relative position information.

[0127] Furthermore, the label pair generating unit includes:

[0128] The label pair generation subunit is used to identify labels with adjacent positions, select two adjacent labels in sequence, and generate the label pair.

[0129] Furthermore, the relative position information includes relative distance and / or relative direction.

[0130] Furthermore, the history label is not blocked.

[0131] Furthermore, the closed-loop detection module 430 includes:

[0132] A standard feature data generation submodule, used to determine the standard feature data corresponding to the target feature data in a preset matching library;

[0133] The closed-loop detection submodule is used to perform closed-loop detection on the movement path of the robot according to the matching situation between the target feature data and the standard feature data.

[0134] Furthermore, the standard feature data generation submodule includes:

[0135] A standard tag group generating unit, used for generating a standard tag group according to the identification information of each tag corresponding to the target feature data;

[0136] The first standard feature data generating unit is used to generate the standard feature data according to the standard position information of each tag in the standard tag group.

[0137] Furthermore, the closed-loop detection module 430 further includes a preset matching library construction submodule, wherein the preset matching library construction submodule includes:

[0138] A standard tag group construction unit, used to construct at least one standard tag group according to the identification information of each standard tag;

[0139] A second standard feature data generating unit, configured to generate the standard feature data according to the position information of each standard tag in the standard tag group;

[0140] The preset matching library construction unit is used to construct the preset matching library according to at least one set of the standard feature data.

[0141] Furthermore, the feature data generation submodule includes:

[0142] A target feature data construction unit is used to sequentially form target tag connection lines of connection line relationship for tags with adjacent identification positions, and generate target feature data based on obtaining connection line attributes of the connection lines of the position feature points of the target tags; wherein the connection line attributes include position information and direction information;

[0143] Accordingly, the closed-loop detection submodule includes:

[0144] The closed-loop detection unit is used to generate standard feature data according to the second connection line attributes of the position feature points of the two target tags; match the connection line attributes of each target tag connection line in the target feature data with the connection line attributes of each standard tag connection line in the standard feature data, and perform closed-loop detection on the movement path of the robot based on the matching results.

[0145] The closed-loop detection device provided in the embodiments of the present application can execute the closed-loop detection method provided in any embodiment of the present application, and has the corresponding functional modules and beneficial effects of the execution method.

[0146] Embodiment 5

[0147] Figure 6 A schematic diagram of the structure of a robot provided in Example 5 of the present application is shown in FIG. Figure 6 As shown, the robot includes a processor 510 , a memory 520 , an input device 530 , and an output device 540 .

[0148] The number of processors 510 in the robot can be one or more. Figure 6 A processor 510 is taken as an example; the processor 510, the memory 520, the input device 530 and the output device 540 in the robot can be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.

[0149] The input device 530 is used to receive the path data collected by the robot.

[0150] The output device 540 is used to output the closed-loop detection result.

[0151] The processor 510 can identify the current label of the robot at the current position according to the path data input by the input device 530; it can also generate target feature data according to the position information of the current label; it can also perform closed-loop detection on the movement path of the robot according to the target feature data; it can also transmit the closed-loop detection result to the output device 540.

[0152] The memory 520, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the closed-loop detection method in the embodiment of the present application (for example, the tag recognition module 410, the feature data generation module 420 and the closed-loop detection module 430 in the closed-loop detection device). The processor 510 executes various functional applications and data processing of the device by running the software programs, instructions and modules stored in the memory 520, that is, implements the above-mentioned closed-loop detection method.

[0153] The memory 520 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. (such as the location information of the current tag, the location information of the historical tag, the target feature data, the standard feature data and the preset matching library in the above embodiment). In addition, the memory 520 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 520 may further include a memory remotely arranged relative to the processor 510, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network and a combination thereof.

[0154] The input device 530 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the device. The output device 540 may include a display device such as a display screen.

[0155] Embodiment 6

[0156] Embodiment 6 of the present application further provides a storage medium containing computer executable instructions, wherein the computer executable instructions are used to execute a closed-loop detection method when executed by a computer processor, the method comprising:

[0157] Identify the current tag of the robot at its current location;

[0158] Generate target feature data according to the position information of the current tag;

[0159] A closed-loop detection is performed on the movement path of the robot according to the target feature data.

[0160] Through the above description of the implementation method, the technicians in the relevant field can clearly understand that the present application can be implemented with the help of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application can be essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0161] It is worth noting that in the embodiment of the above-mentioned closed-loop detection device, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application.

[0162] Note that the above are only preferred embodiments of the present application and the technical principles used. Those skilled in the art will understand that the present application is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present application. Therefore, although the present application is described in more detail through the above embodiments, the present application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the appended claims.

Claims

1. A closed-loop detection method, characterized in that: include: Identify the current tag of the robot at its current location; On the moving path of the robot and before passing the position of the current tag, the robot recognizes at least one historical tag; generating at least one set of label pairs according to the current label and the at least one historical label; According to the movement path of the robot, the position information of each tag in the tag pair is identified during the movement process to generate relative position information; wherein the relative position information is the relative position information between the position information of each tag in the tag pair; the position information of the tag includes coordinate information and angle information; the coordinate information refers to the coordinate information in the robot coordinate system, and the angle information refers to the inclination angle information of the tag and the coordinate plane; Generating target feature data according to at least one of the relative position information, including: sequentially forming target tag connection lines of connection line relationship with tags whose identification positions are adjacent, and generating the target feature data based on connection line attributes of the connection lines of the position feature points of the target tags; wherein the connection line attributes include position information and direction information; The target feature data is matched with the scene feature data stored in the database. If the matching results are consistent, the corresponding matching scene label is output, and the current label identifier is matched with the identifier of the matching scene label. According to the matching results, the movement path of the robot is closed-loop detected.

2. The method according to claim 1, characterized in that: The step of generating at least one set of tag pairs according to the current tag and the at least one historical tag includes: Numbering the current tag and the at least one historical tag according to their recognition order; Two labels with adjacent numbers are selected in sequence to generate the label pair.

3. The method according to claim 1, characterized in that: The relative position information includes relative distance and / or relative direction.

4. The method according to claim 1, characterized in that: Also includes: Determine standard feature data corresponding to the target feature data in a preset matching library; According to the matching condition between the target feature data and the standard feature data, and according to the matching condition between the tag identifier corresponding to the target feature data and the tag identifier corresponding to the standard feature data, a closed-loop detection is performed on the moving path of the robot.

5. The method according to claim 4, characterized in that The preset matching library is constructed in the following manner: Constructing at least one standard tag group according to the identification information of each standard tag; Generate the standard feature data according to the position information of each standard tag in the standard tag group during the movement of the robot's movement path; The preset matching library is constructed according to at least one set of the standard feature data.

6. The method according to claim 1, characterized in that Also includes: storing the second connection line attributes of the position feature points of the two target tags to generate standard feature data; The target feature data is matched with the standard feature data, and the tag identifier corresponding to the target feature data is matched with the tag identifier corresponding to the standard feature data, and based on the matching result, a closed-loop detection is performed on the movement path of the robot, and the position information of the target tag is determined according to a predetermined rule.

7. A closed-loop detection device, characterized in that: include: The tag recognition module is used to recognize the current tag of the robot at the current position; A feature data generation module, including a feature data generation submodule; The feature data generation submodule includes: a tag pair generation unit, a position information generation unit and a feature data generation unit; The label pair generating unit is used for generating at least one set of label pairs according to the current label and the at least one historical label when the robot recognizes at least one historical label on the moving path of the robot and before passing the position of the current label; The position information generating unit is used to identify the position information of each tag in the tag pair according to the movement path of the robot during movement, and generate relative position information; wherein the relative position information is the relative position information between the position information of each tag in the tag pair; the position information of the tag includes coordinate information and angle information; the coordinate information refers to the coordinate information in the robot coordinate system, and the angle information refers to the inclination angle information of the tag and the coordinate plane; The feature data generating unit is used to generate target feature data according to at least one relative position information; The feature data generating unit is specifically used to sequentially form target tag connecting lines of connecting line relationship with tags whose identification positions are adjacent, and generate the target feature data based on the connecting line attributes of the connecting line of the position feature points of the target tags; wherein the connecting line attributes include position information and direction information; The closed-loop detection module is used to match the target feature data with the scene feature data stored in the database. If the matching results are consistent, the corresponding matching scene label is output, and the current label identifier is matched with the identifier of the matching scene label. According to the matching results, the movement path of the robot is closed-loop detected.

8. A robot, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement a closed-loop detection method as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a closed-loop detection method as described in any one of claims 1 to 6 is implemented.

Images