Method for improving positioning precision of robot and storage medium

Abstract

The invention discloses a method for improving the positioning precision of a robot and a storage medium. The method for improving the positioning precision of the robot comprises the steps that a current area where the robot is located is determined according to current pose information of the robot; selecting a corresponding sensor for data acquisition according to the area type of the current area to obtain sensor data; and in response to the abnormal data in the sensor data, screening the abnormal data in the sensor data, and carrying out positioning processing according to the screened sensor data. According to the scheme, abnormal data can be eliminated, and the positioning precision of the robot is improved.

Description

Technical Field

[0001] The present application relates to the field of robotics technology, and in particular to a method for improving the positioning accuracy of a robot and a storage medium. Background Art

[0002] With the rapid development of intelligence, mobile robots are widely used in various industries, such as warehousing and logistics, housekeeping and cleaning, etc. For mobile robots, having accurate positioning and docking methods is of utmost importance.

[0003] For example, in a warehousing scenario, the mobile robot needs to accurately reach the target point of the docking station before it can successfully pick up and place items on the docking station.

[0004] However, the environmental information in the warehousing scene changes frequently, and as the robot works for a long time, it may collide with the sensors on the robot and / or the docking marks on the docking platform (machine), thereby affecting the positioning and docking accuracy of the robot. Summary of the invention

[0005] The present application at least provides a method, an apparatus, a device and a computer-readable storage medium for improving the positioning accuracy of a robot.

[0006] The first aspect of the present application provides a method for improving the positioning accuracy of a robot, comprising: determining the current area where the robot is located according to the current posture information of the robot; selecting a corresponding sensor to collect data according to the area type of the current area to obtain sensor data; in response to the presence of abnormal data in the sensor data, filtering and processing the abnormal data in the sensor data to obtain filtered sensor data; and performing positioning processing based on the filtered sensor data.

[0007] In one embodiment, determining the current area where the robot is located based on the current posture information of the robot includes: determining whether the robot is in a laser recognition area based on an acquired prior map and the current posture information; selecting a corresponding sensor for data collection based on the area type of the current area to obtain sensor data includes: in response to the area type being a laser recognition area, respectively acquiring corresponding laser point cloud data and odometer data based on a current timestamp and a historical timestamp; the historical timestamp is earlier than the current timestamp.

[0008] In one embodiment, in response to the area type being a laser recognition area, after obtaining corresponding laser point cloud data and odometer data according to the current timestamp and the historical timestamp, respectively, the method includes: determining a first laser pose corresponding to the current timestamp according to the laser point cloud data corresponding to the current timestamp, and determining a second laser pose corresponding to the historical timestamp according to the laser point cloud data corresponding to the historical timestamp; determining a first odometer pose corresponding to the current timestamp according to the odometer data corresponding to the current timestamp, and determining a second odometer pose corresponding to the historical timestamp according to the odometer data corresponding to the historical timestamp; performing abnormality judgment based on the first laser pose, the second laser pose, the first odometer pose, and the second odometer pose to obtain the abnormal data.

[0009] In one embodiment, the abnormality judgment is performed based on the first laser posture, the second laser posture, the first odometer posture and the second odometer posture to obtain the abnormal data, including: determining a target posture difference based on a laser posture difference between the first laser posture and the second laser posture, and an odometer posture difference between the first odometer posture and the second odometer posture; in response to the target posture difference being greater than a preset posture difference threshold, determining the first laser posture as the abnormal data.

[0010] In one embodiment, the robot includes a left camera and a right camera respectively arranged according to the driving direction of the robot, and the corresponding sensor is selected according to the area type of the current area to collect data to obtain sensor data, including: in response to the area type being a camera recognition area, obtaining a left identification code recognized by the left camera and a right identification code recognized by the right camera; the camera recognition area includes the left identification code and the right identification code respectively arranged on both sides of the camera recognition area; and according to the number of identification codes recognized by the camera on each side, determining the identification code recognized by the camera on one side as the sensor data.

[0011] In one embodiment, after obtaining the left identification code recognized by the left camera and the right identification code recognized by the right camera in response to the area type being a camera identification area, the method further includes: obtaining a left pitch angle obtained by the left camera recognizing the left identification code, and a right pitch angle obtained by the right camera recognizing the right identification code; in response to a pitch angle deviation between the left pitch angle and the right pitch angle being greater than a preset pitch angle deviation threshold, and a deviation time during which the pitch angle deviation is greater than the preset pitch angle deviation threshold being greater than a preset deviation time threshold, generating a deviation prompt message.

[0012] In one embodiment, after determining the identification code recognized by the camera on one side as the sensor data based on the number of identification codes recognized by the camera on each side, the method further includes: comparing the current number of identification codes in the current sensor data with the number of historical identification codes in the historical sensor data; the collection time of the historical sensor data is earlier than the current sensor data; in response to the difference in quantity between the current number of identification codes and the number of historical identification codes being greater than a preset number difference threshold, determining the current sensor data as the abnormal data.

[0013] In one embodiment, after determining the identification code recognized by the camera on one side as the sensor data based on the number of identification codes recognized by the camera on each side, the method further includes: obtaining the identification distance between the robot and the current identification code in the sensor data; judging whether the sensor data is abnormal based on the identification distance, the number of identification codes currently recognized and a preset number threshold; the identification distance is positively correlated with the preset number threshold.

[0014] In one embodiment, the robot also includes a left photoelectric sensor and a right photoelectric sensor respectively set according to the driving direction of the robot. After the response that the area type is a camera recognition area, the method also includes: obtaining the left photoelectric distance recognized by the left photoelectric sensor and the right photoelectric distance recognized by the right photoelectric sensor, and obtaining the left identification distance between the robot and the current left identification code and the right identification distance between the robot and the current right identification code through the left camera and the right camera; in response to the distance difference between the left photoelectric distance and the left identification distance being greater than a preset distance difference threshold, the current left identification code is determined as the abnormal data; or, the distance difference between the right photoelectric distance and the right identification distance is greater than the preset distance difference threshold, the current right identification code is determined as the abnormal data.

[0015] The second aspect of the present application provides a device for improving the positioning accuracy of a robot, comprising: an area determination module, used to determine the current area where the robot is located according to the current posture information of the robot; a data acquisition module, used to select a corresponding sensor for data acquisition according to the area type of the current area to obtain sensor data; an abnormal screening module, used to screen and process the abnormal data in the sensor data in response to the presence of abnormal data in the sensor data to obtain screened sensor data; and a positioning module, used to perform positioning processing based on the screened sensor data.

[0016] A third aspect of the present application provides an electronic device, including a memory and a processor, wherein the processor is used to execute program instructions stored in the memory to implement the above-mentioned method for improving robot positioning accuracy.

[0017] A fourth aspect of the present application provides a computer-readable storage medium having program instructions stored thereon, which implement the above-mentioned method for improving robot positioning accuracy when the program instructions are executed by a processor.

[0018] The above scheme obtains the current posture information of the robot and determines the current area where the robot is located based on the current posture information. Since different positioning strategies can be set when the robot is in different areas, the corresponding sensor can be selected according to the area type of the current area for data collection to obtain corresponding sensor data. It is determined whether there is abnormal data in the sensor data. In response to the presence of abnormal data in the sensor data, the abnormal data in the sensor data is filtered and processed to obtain the filtered sensor data. Positioning processing is performed based on the filtered sensor data, thereby avoiding the interference of abnormal data on the positioning process of the robot and improving the positioning accuracy of the robot.

[0019] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present application and are used together with the specification to illustrate the technical solution of the present application.

[0021] Figure 1 is a flow chart of an exemplary embodiment of a method for improving robot positioning accuracy of the present application;

[0022] Figure 2 is a schematic diagram of an exemplary machine platform in the method for improving robot positioning accuracy of the present application;

[0023] Figure 3 This is a schematic diagram of an exemplary application scenario of the method for improving robot positioning accuracy of the present application;

[0024] Figure 4 is a schematic diagram of an exemplary identification code in the method for improving robot positioning accuracy of the present application;

[0025] Figure 5 is an exemplary camera calibration schematic diagram in the method for improving robot positioning accuracy of the present application;

[0026] Figure 6 is an exemplary photoelectric verification schematic diagram in the method for improving robot positioning accuracy of the present application;

[0027] Figure 7 is a block diagram of a device for improving robot positioning accuracy shown in an exemplary embodiment of the present application;

[0028] Figure 8 It is a structural schematic diagram of an embodiment of the electronic device of the present application;

[0029] Fig. 9 It is a structural diagram of an embodiment of a computer-readable storage medium of the present application. DETAILED DESCRIPTION

[0030] The scheme of the embodiment of the present application is described in detail below in conjunction with the drawings of the specification.

[0031] In the following description, for the purpose of explanation rather than limitation, specific details such as specific system structures, interfaces, and technologies are provided to facilitate a thorough understanding of the present application.

[0032] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the objects associated before and after are in an "or" relationship. In addition, "many" in this article means two or more than two. In addition, the term "at least one" in this article means any combination of at least two of any one or more of a plurality of, for example, including at least one of A, B, and C, can mean including any one or more elements selected from the set consisting of A, B, and C.

[0033] It is understandable that the method for improving robot positioning accuracy of the present application can be applied to various robots on the market that meet the configuration requirements of the present application. Therefore, the present application does not limit the specific application scenarios and the types of robots targeted (for example, it can be applied to AGV (Automated Guided Vehicle), AMR (Autonomous Mobile Robot), drone robots, etc.). The following text of the present application mainly uses robots in a warehousing environment as an example for explanation.

[0034] These robots usually perform high-precision positioning and movement in a two-dimensional grid map generated based on the warehouse environment. The environment (map) usually includes elements such as shelves, workbenches, building support frames, and goods to be moved. The robots can use data such as odometers, laser point clouds, and prior maps to perform autonomous positioning and navigation.

[0035] See also Figure 1 , Figure 1This is a flowchart of an exemplary embodiment of the method for improving the positioning accuracy of a robot of the present application. The method of the present application can be applied to a robot, or to a software system that is in communication with the robot, which is not limited here. Specifically, it can include the following steps:

[0036] Step S110, determining the current area where the robot is located according to the current position information of the robot.

[0037] The current position and posture information of the robot represents the current position and posture of the robot. The current area of ​​the robot can be determined according to the current position of the robot, and the posture of the robot in the current area can be determined according to the current posture of the robot.

[0038] It is understandable that the method of determining the current position information of the robot may include but is not limited to matching the data such as laser point cloud and / or odometer with the prior map in the current scene, and then determining which area of ​​the prior map the robot is in. For details, please refer to the principle of positioning the robot according to the prior map, which will not be elaborated here.

[0039] Exemplarily, for example, the laser point cloud data currently collected by the robot may be matched with the laser point cloud data in the prior map to thereby determine the current area of ​​the robot in the prior map.

[0040] Step S120: Select a corresponding sensor to collect data according to the area type of the current area to obtain sensor data.

[0041] It should be noted that the prior map can be pre-divided into multiple areas, and the robot can have different operating methods when in different areas. Different operating methods may correspond to the selection of different sensors for data collection, and then obtain one or more sensor data, which is not limited here.

[0042] For example, when a robot is moving on a normal road, it usually uses a laser radar sensor to collect laser point cloud data in the current area, and then combines it with a priori maps to achieve positioning and navigation functions. When the robot moves to the docking point of the machine, it may turn on the image sensor (camera) to collect image data, and achieve precise docking between the robot and the machine based on the image data and laser point cloud data, so that the robot can accurately reach the target point and correctly perform subsequent tasks (such as picking up and placing items placed on the machine, etc.).

[0043] In this application, the basic settings of the robot may include motion controllers, motors, batteries, embedded computers, odometers, etc., and on this basis, multiple sensors (including but not limited to laser radars, cameras, photoelectric sensors, etc.) may also be provided. The acquisition frequency of the odometer data of this application may be greater than the acquisition frequency of the laser radar data, and the acquisition frequency of the laser radar data may be greater than the acquisition frequency of the image data. Therefore, during the operation of the robot, the corresponding sensor can be selected according to the area type of the current area for data acquisition to obtain the corresponding sensor data.

[0044] It should be noted that in the specific implementation process of the present application, the corresponding sensor can be selected to be turned on and data can be collected according to the type of the current area where the robot is located; or each sensor can be turned on and data can be collected during the entire operation of the robot, and the corresponding sensor data can be selected for use and processing according to the type of the current area where the robot is located. There is no limitation here.

[0045] Step S130 , in response to the presence of abnormal data in the sensor data, filtering the abnormal data in the sensor data to obtain filtered sensor data.

[0046] In combination with the above steps, if abnormal data is detected in the collected sensor data, it is necessary to filter and process the abnormal data in the sensor data to obtain filtered sensor data, so as to avoid being affected by the abnormal data in the process of controlling the robot for positioning navigation or performing other tasks according to the sensor data, thereby reducing the robot's operating accuracy.

[0047] Exemplarily, the screening of abnormal data may include but is not limited to screening out the abnormal data from the sensor data and retaining the remaining normal data as the screened sensor data. The screened abnormal data may be directly deleted or stored in other locations so that subsequent abnormal analysis can be performed based on the abnormal data, which will not be described in detail here.

[0048] Step S140: performing positioning processing according to the filtered sensor data.

[0049] It should be noted that during the robot's positioning and navigation process, the robot can be repositioned according to the sensor data obtained each time (or the sensor data obtained according to a preset periodic frequency) and / or the target point that the robot needs to go to can be repositioned. In this way, the robot's driving path is iteratively covered (or iteratively updated) by overwriting the old positioning data with the new positioning data, which can improve the robot's positioning and navigation accuracy.

[0050] In conjunction with the above steps, in the specific implementation process of the present application, on the one hand, after filtering the abnormal data, positioning processing is performed based on the filtered sensor data, thereby improving the accuracy of robot positioning and navigation. On the other hand, it is also possible to identify based on the sensor data before filtering, and remove the recognition results of the abnormal data in the sensor data (abnormal recognition results), and not use the abnormal recognition results for path coverage.

[0051] It should also be noted that if the currently identified sensor data is abnormal data, since the abnormal data is not used for positioning and navigation, the normal results previously determined based on the normal data can still be used for positioning and navigation until the next normal result is identified and the path is covered. This will not be elaborated here.

[0052] It can be seen that the present application obtains the current posture information of the robot, and determines the current area where the robot is located based on the current posture information. Since different positioning strategies can be set when the robot is in different areas, the corresponding sensor can be selected according to the area type of the current area for data collection to obtain corresponding sensor data. It is determined whether there is abnormal data in the sensor data. In response to the presence of abnormal data in the sensor data, the abnormal data in the sensor data is screened and processed to obtain the screened sensor data. Positioning processing is performed based on the screened sensor data, thereby avoiding interference of abnormal data on the positioning process of the robot and improving the positioning accuracy of the robot.

[0053] Based on the above embodiments, the present embodiment of the application describes the steps of determining the current area of ​​the robot according to the current posture information of the robot and selecting the corresponding sensor for data collection according to the area type of the current area to obtain sensor data. Specifically, the method of this embodiment includes the following steps:

[0054] Determine whether the robot is in the laser recognition area based on the acquired prior map and current posture information; in response to the area type being a laser recognition area, obtain corresponding laser point cloud data and odometer data based on the current timestamp and historical timestamp respectively; the historical timestamp is earlier than the current timestamp.

[0055] For example, reference may be made to Figure 2 As shown, Figure 2This is an exemplary schematic diagram of a machine in the method of improving the positioning accuracy of the robot in the present application. Among them, a reflective mark is set at the machine (the reflective mark can be a rectangle or other shapes, which is not limited here). Since the laser radar can obtain more obvious point cloud data when it is irradiated on the reflective mark, the reflective mark can be used to improve the accuracy of laser recognition and positioning. Among them, the distance between the leftmost reflective mark to the rightmost reflective mark can be set to at least 1m to ensure the angle accuracy of the recognition. In addition, three reflective marks can be installed, and the width of each reflective mark can be 4cm to 6cm. The above setting specifications can be flexibly adjusted according to the specific application scenario. This is only an example and is not limited.

[0056] As another example, reference may also be made to Figure 3 As shown, Figure 3 This is a schematic diagram of an exemplary application scenario in the method of improving robot positioning accuracy of the present application. Among them, the laser recognition area or laser recognition point can be pre-divided in the prior map. When the robot moves to the laser recognition area or laser recognition point, it is necessary to perform laser recognition on the reflective mark set on the machine to obtain the corresponding laser point cloud data, and the position information can be determined according to the laser point cloud data, and then the positioning navigation and other tasks can be performed.

[0057] Specifically, the laser radar timestamp (current timestamp) of the current reflective mark recognized by the laser radar can be obtained and recorded as tl k , and record the laser recognition result as T lmk According to the lidar timestamp, find the distance from the current timestamp t in the storage queue of the odometer data. lk The two closest frames of odometer data are timestamped as t ok-1 and t ok , and the robot pose corresponding to the two frames of odometer data is recorded as T ok-1 and T ok Among them, it should be noted that t ok-1 and t ok and t lk The temporal relationship between them can be t ok-1 <t ok <t lk , or t ok-1 <t lk <t ok , or t lk <t ok-1 <t ok , which is not limited here. Thus, the current timestamp t can be obtained lk The corresponding robot pose T bk . For example, one of the mathematical expressions may be:

[0058]

[0059] In addition, the previous laser recognition result is determined based on the historical timestamp before the current timestamp. For example, the recognition result corresponding to the previous frame of laser data of the current frame of laser data is obtained, and the recognition result obtained by recognizing the previous frame of laser data is recorded as T lmk-1 Similarly, the odometer data corresponding to the historical timestamp can be determined according to the above method as T bk-1 . Thus, the sensor data corresponding to the laser recognition area obtained by the robot is obtained.

[0060] Based on the above embodiment, the embodiment of the present application describes the steps after, in response to the area type being a laser recognition area, the corresponding laser point cloud data and odometer data are acquired according to the current timestamp and the historical timestamp. Specifically, the method of this embodiment includes the following steps:

[0061] Determine the first laser pose corresponding to the current timestamp according to the laser point cloud data corresponding to the current timestamp, and determine the second laser pose corresponding to the historical timestamp according to the laser point cloud data corresponding to the historical timestamp; determine the first odometer pose corresponding to the current timestamp according to the odometer data corresponding to the current timestamp, and determine the second odometer pose corresponding to the historical timestamp according to the odometer data corresponding to the historical timestamp; perform abnormal judgment based on the first laser pose, the second laser pose, the first odometer pose, and the second odometer pose to obtain abnormal data.

[0062] Combined with the above embodiment, according to the current timestamp t lk The corresponding laser point cloud data is laser recognized to determine the current timestamp t lk The corresponding first laser pose T lmk Similarly, according to the historical timestamp (for example, t lk-1 ) to perform laser recognition on the corresponding laser point cloud data, and determine the second laser pose T corresponding to the historical timestamp lmk-1 Through the method provided in the above embodiment, according to the current timestamp t lk Determined t ok-1 and t ok The odometer data can determine t lk The corresponding first odometer pose T bk Similarly, it is also possible to use the historical timestamp t lk-1 Determine the corresponding first odometer pose T bk-1 .

[0063] Therefore, according to T lmk , T lmk-1 , T bk and Tbk-1 An abnormality judgment is performed to determine whether there is abnormal data in the sensor data obtained above.

[0064] On the basis of the above embodiments, the embodiment of the present application describes the steps of performing abnormality judgment according to the first laser posture, the second laser posture, the first odometer posture and the second odometer posture to obtain abnormal data. Specifically, the method of this embodiment includes the following steps:

[0065] A target pose difference is determined based on a laser pose difference between the first laser pose and the second laser pose, and an odometer pose difference between the first odometer pose and the second odometer pose; in response to the target pose difference being greater than a preset pose difference threshold, the first laser pose is determined as abnormal data.

[0066] Combined with the above embodiment, the method of the above embodiment can obtain the first laser posture T lmk , the second laser pose T lmk-1 , the first odometer posture T b k and second odometry pose T bk-1 .

[0067] Furthermore, according to the first laser posture T lmk and the second laser pose T lmk-1 Calculate the laser pose difference (angle difference) deltal between the two poses; according to the first odometer pose T bk and the second odometer pose T bk-1 Calculate the odometer pose difference (angle difference) delta2 between the two poses. Determine the target pose difference based on the two pose differences (delta1 and delta2) (for example, calculate the difference between delta1 and delta2). If the target pose difference is greater than the preset pose difference threshold th1 (for example, th1 = 2° can be set), it can be determined that at the current timestamp t lk The acquired LiDAR data is abnormal data, and / or it is determined based on t lk Laser recognition result T obtained from laser radar data lmk This is abnormal data. Therefore, the laser recognition result can be discarded and not used for positioning and navigation.

[0068] Furthermore, an abnormality counter may be set during the specific implementation of this embodiment, and the abnormality counter will accumulate the count of each abnormal data detected, and reset (clear) if normal data is detected. If the above-mentioned type of abnormal data is detected multiple times (greater than or equal to the preset number of times), all the obtained laser recognition results may be screened and eliminated.

[0069] Optionally, if the laser recognition result T corresponding to the current timestamp lmk The laser recognition result T corresponding to the historical timestamp lmk-1 If the angle difference between them is greater than the preset angle threshold th2 (for example, th2=5° can be set, th1 and th1 can be the same or different, which is not limited here), the laser radar data obtained at the current timestamp can be judged as abnormal data and eliminated in the same way. This type of judgment process and the above judgment process can be implemented simultaneously or one of them can be implemented at will, which is not limited here. Since the length of the reflective mark recognized by the laser radar is relatively long, the docking angle calculated using the laser data is very accurate. Therefore, when using the angle to eliminate abnormal data in the subsequent process, the angle value represented by the laser radar recognition result can be more trusted.

[0070] Based on the above embodiment, the embodiment of the present application describes the steps of selecting a corresponding sensor to collect data according to the area type of the current area to obtain sensor data. The robot includes a left camera and a right camera respectively arranged on the left and right sides of the robot's driving direction. Specifically, the method of this embodiment includes the following steps:

[0071] In response to the area type being a camera recognition area, a left identification code recognized by a left camera and a right identification code recognized by a right camera are obtained; the camera recognition area includes a left identification code and a right identification code respectively arranged on both sides of the camera recognition area; and according to the number of identification codes recognized by the camera on each side, the identification code recognized by the camera on one side is determined as sensor data.

[0072] Continue to combine Figure 2 To explain, Figure 2 There are identification codes (identification code belts) on both sides of the machine, which can be set according to the identification code, ArUco code, etc., and are also one of the identifications used by the robot to achieve precise positioning. Figure 3 For explanation, according to the robot's driving direction, a left camera and a right camera are provided on the left and right sides of the robot (there can be one or more left and right cameras, which is not limited here). When the robot drives to the camera recognition area, the left camera can be turned on to identify the left identification code on the left side, and the right camera can be turned on to identify the right identification code on the right side.

[0073] It can be understood that the present application can calculate the position and posture of the camera relative to the identification code based on the identification code recognized by the camera and use the PnP (Perspective-n-Point) method to determine the position and posture of the robot. For details, please refer to the existing robot visual positioning method, which will not be elaborated here.

[0074] For example, the setting method of the identification code can refer to the Figure 4 , Figure 4This is an exemplary identification code schematic diagram in the method of improving robot positioning accuracy of the present application. It shows the setting method of the left identification code and the right identification code. A reference identification code (reference code) can be set in each of the left identification code and the right identification code. The left and right reference codes can be aligned with the target point (for example, the left and right reference codes and the target point are on the same straight line), so that when the robot recognizes the reference code, it can determine the target point and perform tasks such as picking up and placing goods at the target point. This is one of the optional setting methods for the warehousing environment, and is not specifically limited.

[0075] For example, after the robot enters the camera recognition area, the left and right camera recognition can be turned on (the laser radar-based reflective sign recognition can continue) to verify whether the camera recognition angle is correct. Figure 4 As shown, the left and right identification codes of the present application may include one or more, the number of left and right identification codes is the same and the IDs of the left and right identification codes correspond to the same position based on the identification codes, which is not limited here. During the movement of the robot, the number of identification codes recognized by the left and right cameras can be determined respectively, so the data collected by the camera on the left and right sides with a larger number of identification codes can be used as sensor data for subsequent robot positioning and navigation to perform path coverage (for example, if the number of left identification codes recognized by the left camera is greater than the number of right identification codes recognized by the right camera, positioning and navigation processing is performed based on the image data collected by the left camera). Among them, the selection of left data or right data can be fixed or flexibly changed according to actual conditions, which is not limited here (for example, if the number of right identification codes recognized by the right camera is greater than the number of left identification codes recognized by the left camera, it is switched to positioning and navigation processing based on the image data collected by the right camera).

[0076] It should be noted that, in order to ensure the consistency of the recognition results of the cameras on both sides, the relative positions of the left and right cameras of the present application may be pre-calibrated. Figure 5 As shown, Figure 5 This is an exemplary camera calibration diagram in the method of improving robot positioning accuracy of the present application. For example, let the left camera be C0, the right camera be C1, two completely parallel left and right identification codes (left and right reference codes) are known, and the distance between the two identification codes is known to be The relative position of the left and right cameras is The mathematical expression of can be:

[0077]

[0078] in, is the external parameter of the right camera relative to the left camera; It is the position and posture of the left camera relative to the left reference code calculated by the pnp method when the left camera scans the reference code Vm0; It is the position and posture of the right camera relative to the right reference code Vm1 calculated by the pnp method when the right camera scans the right reference code Vm1.

[0079] Furthermore, since the camera of the robot of the present application has been calibrated, when the left camera recognizes the left identification code, the position of the left camera relative to the left identification code is recorded as T cvl ; When the right camera recognizes the same ID as the left camera, the position of the right camera relative to the right ID is recorded as T cvr ; Since the left and right cameras have been calibrated, T cvl =T cvr .

[0080] Therefore, the coordinates T of the target point determined by the camera data can be obtained ctm The mathematical expression is:

[0081] T ctm =T bm *T cb *T c0cx *T cvl

[0082] Among them, T c0cx is the relative position of the current identification code scanned by the camera relative to the reference identification code (pre-set and directly obtainable), c0 refers to the reference identification code, T cb is the external parameter of the camera relative to the robot body (pre-set and directly accessible), T bm is the current robot position determined during camera recognition. The robot reaches the target position T ctm It can complete the docking with the machine and perform operations such as picking and placing goods.

[0083] Optionally, at the same time, when the target point T determined by the camera identifying the identification code ctm The target point T is determined by laser recognition of reflective markings ltm When the angle difference between them is greater than a threshold value th3 (for example, 1°, th1, th2 and th3 may be the same or different, and are not limited here), the camera recognition result at that moment may be discarded.

[0084] Based on the above embodiment, the embodiment of the present application describes the steps after obtaining the left identification code recognized by the left camera and the right identification code recognized by the right camera in response to the area type being the camera identification area. Specifically, the method of this embodiment includes the following steps:

[0085] A left pitch angle obtained by the left camera recognizing the left identification code and a right pitch angle obtained by the right camera recognizing the right identification code are obtained; in response to a pitch angle deviation between the left pitch angle and the right pitch angle being greater than a preset pitch angle deviation threshold, and a deviation time during which the pitch angle deviation is greater than the preset pitch angle deviation threshold being greater than a preset deviation time threshold, a deviation prompt message is generated.

[0086] In conjunction with the above-mentioned embodiment, when the left and right cameras of the robot recognize the left and right identification codes, it is also possible to determine whether the left and right cameras and / or the left and right identification codes of the robot are set incorrectly (for example, the left and right cameras may tilt as the robot continues to move, or the identification codes may tilt due to external impacts). The detection process of this embodiment may be performed throughout the entire camera recognition process or selectively, which is not limited here.

[0087] Exemplarily, when the left camera recognizes the left identification code, the left pitch angle pitch1 of the left camera can be calculated. Similarly, when the right camera recognizes the right identification code, the right pitch angle pitch2 of the right camera can be calculated. The pitch angle deviation between the left pitch angle pitch1 and the right pitch angle pitch2 is calculated. In response to the pitch angle deviation being greater than a preset pitch angle deviation threshold (e.g., 2°), and the duration of this deviation (deviation time) being greater than the preset deviation time threshold, a deviation prompt message can be generated to remind the user to check the identification code and / or camera.

[0088] Another example is that if the recognition result (posture) of the left camera for the left identification code or the recognition result (posture) of the right camera for the right identification code is continuously detected and there is an angular deviation from the laser recognition result (posture) for a long time, a deviation prompt information can be generated in the same way.

[0089] Based on the above embodiment, the embodiment of the present application describes the steps after determining the identification code recognized by the camera on one side as sensor data according to the number of identification codes recognized by the camera on each side. Specifically, the method of this embodiment includes the following steps:

[0090] The current number of identification codes in the current sensor data is compared with the number of historical identification codes in the historical sensor data; the collection time of the historical sensor data is earlier than the current sensor data; in response to the difference in quantity between the current number of identification codes and the number of historical identification codes being greater than a preset number difference threshold, the current sensor data is determined to be abnormal data.

[0091] In conjunction with the above-mentioned embodiments, in the implementation process of this application, in addition to selecting the data on the side with a larger number of identification codes from the left and right sides for positioning and navigation, the number of identification codes currently identified can also be compared with the number of identification codes previously identified (the number of historical identification codes) to determine whether there is an abnormality. Among them, the identification code previously identified refers to the identification code identified earlier than the current moment during the robot's movement to the target point this time, that is, during the robot's movement to the target point this time, the collection time of the historical sensor data is earlier than the collection time of the current sensor data.

[0092] Exemplarily, when covering the target point and / or driving path determined by the last camera recognition result according to the current camera recognition result, the number of identification codes currently scanned by the camera will be judged. If the number of identification codes scanned this time is less than a certain value (equivalent to the number difference between the current number of identification codes and the number of historical identification codes being greater than a preset number difference threshold) than the number of identification codes scanned when the last recognition result was covered, for example, less than 2 or more, the path coverage according to the current camera recognition result can be paused or stopped. Wherein, in the case of limiting the number difference, the recognition distance can also be limited. Because the recognition distance is too close, the camera may only recognize one identification code or even fail to recognize a complete identification code. Therefore, the above judgment process can be implemented when the recognition distance between the camera and the identification code is greater than the preset distance threshold (for example, 10cm). Otherwise, even if the number difference between the current number of identification codes and the number of historical identification codes is greater than the preset number difference threshold, it cannot be determined as abnormal, and the path coverage process can still be performed as usual.

[0093] In another exemplary embodiment, when the camera scans the reference code, in order to ensure the accuracy of the camera's final recognition result, the present application may set more identification codes or double-row identification code bands on the basis of the reference code (for example, see Figure 4 As shown in the figure, the driving direction of the robot and the machine for docking is used as the standard reference direction, and two redundant identification codes are set in front of the reference code. When the camera detects the ID of the reference code, if the number of ArUco codes recognized in the camera field of view is less than a certain identification code number threshold (in Figure 4 In the example scenario, it can be set to 3), then the recognition results containing the reference code can be prohibited from being overwritten.

[0094] Based on the above embodiment, the embodiment of the present application describes the steps after determining the identification code recognized by the camera on one side as sensor data according to the number of identification codes recognized by the camera on each side. Specifically, the method of this embodiment includes the following steps:

[0095] Obtain the recognition distance between the robot and the current identification code in the sensor data; determine whether the sensor data is abnormal based on the recognition distance, the number of identification codes currently recognized, and a preset number threshold; the recognition distance is positively correlated with the preset number threshold.

[0096] It is understandable that in different application scenarios, the robot is at different distances from the identification codes on both sides, which may affect the camera's recognition accuracy for the identification code. For example, for a single ArUco code of the same size, the longer the recognition distance between the camera and the identification code, the lower the recognition accuracy. Therefore, the present application can also limit the number of identification codes that need to be scanned when the path is covered in combination with the recognition distance.

[0097] In conjunction with the above-mentioned embodiment, the identification distance between the robot and the currently identified identification code (current identification code) is obtained, or the identification distance between the camera and the current identification code is obtained, which is not limited here. Whether the sensor data is abnormal is determined based on the identification distance, the number of currently identified identification codes, and the preset number threshold. Among them, the preset number threshold can be positively adjusted according to the identification distance. That is, when the identification distance is farther, more identification codes are required to be currently identified in order to perform the path coverage operation.

[0098] For example, when the recognition distance calculated by the camera is less than 10 cm, at least one identification code is scanned to allow path coverage. When the recognition distance is 10-20 cm, at least two identification codes must be scanned to allow path coverage. When the recognition distance is 20-30 cm, at least three identification codes are scanned to allow path coverage. When the recognition distance is 30-40 cm, the camera needs to scan at least four identification codes to allow path coverage.

[0099] Based on the above embodiment, the embodiment of the present application describes the steps after responding that the area type is a camera recognition area. Among them, the robot of the present application can also be provided with a photoelectric sensor, including a left photoelectric sensor and a right photoelectric sensor respectively provided on the left and right sides of the robot's driving direction. Specifically, the method of this embodiment includes the following steps:

[0100] The left photoelectric distance recognized by the left photoelectric sensor and the right photoelectric distance recognized by the right photoelectric sensor are obtained, and the left identification distance between the robot and the current left identification code and the right identification distance between the robot and the current right identification code are obtained through the left camera and the right camera; in response to the distance difference between the left photoelectric distance and the left identification distance being greater than a preset distance difference threshold, the current left identification code is determined as abnormal data; or, the distance difference between the right photoelectric distance and the right identification distance is greater than the preset distance difference threshold, the current right identification code is determined as abnormal data.

[0101] With reference to the foregoing embodiments, during the implementation of the present application, if the robot enters the camera recognition area, in addition to abnormal data screening based on laser recognition and camera recognition, abnormal data screening can also be performed based on the photoelectric sensors provided on the robot.

[0102] For example, after the robot enters the camera recognition area, the left and right photoelectric sensors on the robot can be turned on at the same time for abnormality verification. Since the detection accuracy of the photoelectric sensor can reach the sub-millimeter level, further verification of the camera recognition result using photoelectric ranging information can further improve the positioning and navigation accuracy of the robot.

[0103] You can refer to Figure 6 As shown, Figure 6 FIG. 1 is an exemplary photoelectric verification diagram of the method for improving the positioning accuracy of the robot in the present application. Figure 6 In the figure, the robot has entered the machine and is ready to dock. The angle in the robot's current posture information is recorded as β, which can be calculated based on the position offset between the identification codes recognized by the left and right cameras. This will not be described in detail here. Figure 6 The mathematical expression of the angle a2 between the robot and the machine platform is:

[0104] a2=Π / 2–β

[0105] In addition, we can also get a1=arctan(c1 / c2). Among them, c1 and c2 are the external parameter distances of the photoelectric sensor data relative to the center of the robot body (based on the docking driving direction) in the vertical direction and the horizontal direction respectively. The mathematical expression of the left photoelectric distance d1 and the right photoelectric distance d2 from the center of the robot body to the identification code tapes on both sides that can be calculated by the left distance l1 and the right distance l2 detected by the photoelectric sensor is:

[0106]

[0107] The left photoelectric distance d1 represents the distance from the center of the vehicle body to the left identification code calculated by the left photoelectric, and the right photoelectric distance d2 represents the distance from the vehicle body to the right identification code calculated by the right photoelectric. When the difference between d1 and d2 is greater than a certain threshold (such as 4cm), it can be considered that the docking deviation between the robot and the machine is too large, and the robot needs to retreat or move sideways to adjust the left and right deviations until the difference between d1 and d2 is less than or equal to the above threshold.

[0108] In addition to determining the distance between the robot and the identification codes on the left and right sides based on the left and right photoelectric sensors, the distance between the robot and the identification codes on the left and right sides needs to be identified by the left and right cameras. The specific calculation method can refer to the existing machine vision technology and will not be described here. The identification code identified by the left camera can calculate the depth value dc1 (left identification distance), and the identification code identified by the right camera can calculate the depth value dc2 (right identification distance).

[0109] In response to the distance difference between the left photoelectric distance d1 and the left identification distance dc1 being greater than the preset distance difference threshold (e.g., 2 cm), the current camera recognition result can be considered abnormal, and the current left identification code or the related data determined based on the current left identification code is determined as abnormal data, and the path coverage is not performed based on this data, thereby ensuring the accuracy of the previous recognition result coverage. Alternatively, if the distance difference between the right photoelectric distance d2 and the right identification distance dc2 is greater than the preset distance difference threshold, the current right identification code is determined as abnormal data.

[0110] On the basis of the above embodiments, the embodiments of the present application also need to explain that two groups of cameras and two groups of photoelectrics can be set on the robot of the present application. Considering that when the robot enters the machine deeper, when the front photoelectric and / or the front camera cannot scan the side identification code, it will switch to the two rear photoelectrics and cameras for identification and detection in time. The method of screening abnormal data is consistent with the screening method of the first two cameras and photoelectrics. The description of the above embodiments can be referred to in the same way, and no limitation is made here.

[0111] In summary, this application combines multiple sensors such as wheel encoders, lidar, photoelectric and cameras. LiDAR is used to identify reflective marks in the laser recognition area to confirm the docking point. At this time, it can be switched to odometer mode for docking. During the docking process, abnormal data is eliminated according to the change in odometer. After entering the camera recognition area, turn on camera recognition, and combine the difference in recognition results of the left and right cameras for the identification codes on the left and right sides, the difference with the laser recognition results, and the number of recognized QR codes to ensure the correctness of the recognition results, thereby ensuring the final precise docking.

[0112] You can also use the preset parallel and aligned identification code strips to calibrate the camera on the robot, and determine the position of the left and right cameras relative to the identification code based on the identification code detected by the left and right cameras to determine the final docking target point. If the difference between the two docking postures is greater than a certain threshold, it can be considered that the recognition result is different or the identification code setting has changed, and a prompt message is generated.

[0113] In addition, the technology of fusion recognition and detection of three sensors, laser, camera and photoelectric, is used to achieve the final high-precision docking requirements. The verification of the three data can use the technology of moving and recognizing to eliminate the cumulative error of the odometer. And because the angle of laser detection is highly accurate, the angle of the laser recognition result is compared with the angle of the camera recognition. If the angle difference is greater than a certain threshold, the abnormal camera recognition result can be eliminated. The angle changes at adjacent moments relative to the reflective mark recognized by the laser and the angle changes at adjacent moments relative to the identification code recognized by the camera are judged with the angle changes of the odometer at the corresponding moments to screen and eliminate abnormal data.

[0114] Furthermore, to ensure that abnormal data can be eliminated when the identification code is dirty or the identification code tape has consistent installation deviation, the robot can be controlled to move to the target point, and the calibration value of the robot during docking can be obtained by statically identifying the reference code. During the process of moving and identifying, if the angle between the camera recognition result and the calibration value is greater than a certain threshold (for example, 1°), the camera recognition result can be considered as abnormal data, and path coverage will not be performed based on this.

[0115] It should be further explained that the execution subject of the method for improving the positioning accuracy of the robot may be a device for improving the positioning accuracy of the robot. For example, the method for improving the positioning accuracy of the robot may be executed by a terminal device or a server or other processing device, wherein the terminal device may be a user equipment (UE), a computer, a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (PDA), a handheld device, a computing device, a vehicle-mounted device, a wearable device, etc. In some possible implementations, the method for improving the positioning accuracy of the robot may be implemented by a processor calling a computer-readable instruction stored in a memory.

[0116] Figure 7 FIG. 1 is a block diagram of a device for improving robot positioning accuracy according to an exemplary embodiment of the present application. Figure 7 As shown, the exemplary device 700 for improving robot positioning accuracy includes: an area determination module 710, a data acquisition module 720, an abnormality screening module 730 and a positioning module 740. Specifically:

[0117] The area determination module 710 is used to determine the current area where the robot is located according to the current position information of the robot.

[0118] The data collection module 720 is used to select a corresponding sensor to collect data according to the area type of the current area to obtain sensor data.

[0119] The abnormal screening module 730 is used to screen the abnormal data in the sensor data in response to the presence of abnormal data in the sensor data to obtain screened sensor data.

[0120] The positioning module 740 is used to perform positioning processing according to the filtered sensor data.

[0121] In this exemplary device for improving the positioning accuracy of a robot, the current position information of the robot is obtained, and the current area where the robot is located is determined based on the current position information. Since different positioning strategies can be set when the robot is in different areas, the corresponding sensor can be selected according to the area type of the current area for data collection to obtain corresponding sensor data. It is determined whether there is abnormal data in the sensor data. In response to the presence of abnormal data in the sensor data, the abnormal data in the sensor data is filtered and processed to obtain the filtered sensor data. Positioning processing is performed based on the filtered sensor data, thereby avoiding interference of abnormal data on the positioning process of the robot, and improving the positioning accuracy of the robot.

[0122] It should be noted that the device provided in the above embodiment and the method provided in the above embodiment belong to the same concept, wherein the specific manner in which each module and unit performs the operation has been described in detail in the method embodiment, and will not be repeated here. In practical applications, the device provided in the above embodiment can distribute the above functions to different functional modules as needed, that is, divide the internal structure of the device into different functional modules to complete all or part of the functions described above, and this is not limited here.

[0123] The functions of each module can be found in the method embodiment for improving robot positioning accuracy, which will not be described in detail here.

[0124] See also Figure 8 , Figure 8 1 is a schematic diagram of the structure of an embodiment of an electronic device of the present application. The electronic device 100 includes a memory 101 and a processor 102, and the processor 102 is used to execute program instructions stored in the memory 101 to implement the steps in any of the above-mentioned methods for improving the positioning accuracy of the robot. In a specific implementation scenario, the electronic device 100 may include but is not limited to: a microcomputer, a server, and in addition, the electronic device 100 may also include a mobile device such as a laptop computer and a tablet computer, which is not limited here.

[0125] Specifically, the processor 102 is used to control itself and the memory 101 to implement the steps in any of the above-mentioned method embodiments for improving the positioning accuracy of the robot. The processor 102 can also be called a CPU (Central Processing Unit). The processor 102 may be an integrated circuit chip with signal processing capabilities. The processor 102 can also be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field-programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. In addition, the processor 102 can be implemented by an integrated circuit chip.

[0126] In this exemplary electronic device, the current position information of the robot is obtained, and the current area where the robot is located is determined based on the current position information. Since different positioning strategies can be set when the robot is in different areas, the corresponding sensor can be selected according to the area type of the current area for data collection to obtain corresponding sensor data. It is determined whether there is abnormal data in the sensor data. In response to the presence of abnormal data in the sensor data, the abnormal data in the sensor data is filtered and processed to obtain the filtered sensor data. Positioning processing is performed based on the filtered sensor data, thereby avoiding interference of abnormal data on the positioning process of the robot and improving the positioning accuracy of the robot.

[0127] See also Fig. 9 , Fig. 9 The computer-readable storage medium 110 stores program instructions 111 that can be executed by a processor, and the program instructions 111 are used to implement the steps in any of the above-mentioned methods for improving the positioning accuracy of a robot.

[0128] In this exemplary storage medium, the current posture information of the robot is obtained by running the program instructions in the storage medium, and the current area where the robot is located is determined according to the current posture information. Since different positioning strategies can be set when the robot is in different areas, the corresponding sensor can be selected according to the area type of the current area for data collection to obtain corresponding sensor data. It is determined whether there is abnormal data in the sensor data. In response to the presence of abnormal data in the sensor data, the abnormal data in the sensor data is screened and processed to obtain the screened sensor data. Positioning processing is performed based on the screened sensor data, thereby avoiding interference of abnormal data on the positioning process of the robot and improving the positioning accuracy of the robot.

[0129] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the method described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0130] The above description of various embodiments tends to emphasize the differences between the various embodiments. The same or similar aspects can be referenced to each other, and for the sake of brevity, they will not be repeated herein.

[0131] In the several embodiments provided in the present application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device implementation described above is only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation, such as units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0132] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (processor) to perform all or part of the steps of each implementation method of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

Claims

1. A method for improving robot positioning accuracy, characterized in that: The method comprises: Determining a current area where the robot is located according to the current position information of the robot; Select a corresponding sensor to collect data according to the area type of the current area to obtain sensor data; In response to the presence of abnormal data in the sensor data, filtering the abnormal data in the sensor data to obtain filtered sensor data; Perform positioning processing based on the filtered sensor data.

2. The method according to claim 1, characterized in that The determining the current area where the robot is located according to the current position information of the robot includes: Determining whether the robot is in a laser recognition area according to the acquired prior map and the current posture information; The step of selecting a corresponding sensor to collect data according to the area type of the current area to obtain sensor data includes: In response to the area type being a laser recognition area, corresponding laser point cloud data and odometer data are respectively acquired according to a current timestamp and a historical timestamp; the historical timestamp is earlier than the current timestamp.

3. The method according to claim 2, characterized in that After, in response to the area type being a laser recognition area, corresponding laser point cloud data and odometer data are acquired according to the current timestamp and the historical timestamp, respectively, the method includes: Determine a first laser pose corresponding to the current timestamp according to the laser point cloud data corresponding to the current timestamp, and determine a second laser pose corresponding to the historical timestamp according to the laser point cloud data corresponding to the historical timestamp; Determine a first odometer posture corresponding to the current timestamp according to the odometer data corresponding to the current timestamp, and determine a second odometer posture corresponding to the historical timestamp according to the odometer data corresponding to the historical timestamp; An abnormality judgment is performed according to the first laser posture, the second laser posture, the first odometer posture, and the second odometer posture to obtain the abnormal data.

4. The method according to claim 3, characterized in that The abnormality judgment is performed according to the first laser posture, the second laser posture, the first odometer posture and the second odometer posture to obtain the abnormal data, including: Determining a target pose difference based on a laser pose difference between the first laser pose and the second laser pose, and an odometer pose difference between the first odometer pose and the second odometer pose; In response to the target posture difference being greater than a preset posture difference threshold, the first laser posture is determined as the abnormal data.

5. The method according to claim 1, characterized in that The robot includes a left camera and a right camera respectively arranged according to the driving direction of the robot, and the corresponding sensor is selected according to the area type of the current area to collect data to obtain sensor data, including: In response to the area type being a camera recognition area, acquiring a left identification code recognized by the left camera and a right identification code recognized by the right camera; the camera recognition area includes the left identification code and the right identification code respectively arranged on both sides of the camera recognition area; According to the number of identification codes recognized by the camera on each side, the identification codes recognized by the camera on one side are determined as the sensor data.

6. The method according to claim 5, characterized in that After acquiring the left identification code recognized by the left camera and the right identification code recognized by the right camera in response to the area type being a camera identification area, the method further includes: Acquire a left pitch angle obtained by the left camera recognizing the left identification code, and a right pitch angle obtained by the right camera recognizing the right identification code; In response to a pitch angle deviation between the left pitch angle and the right pitch angle being greater than a preset pitch angle deviation threshold, and a deviation time during which the pitch angle deviation is greater than the preset pitch angle deviation threshold being greater than a preset deviation time threshold, deviation prompt information is generated.

7. The method according to claim 5, characterized in that After determining the identification code recognized by the camera on one side as the sensor data according to the number of identification codes recognized by the camera on each side, the method further includes: Comparing the number of current identification codes in the current sensor data with the number of historical identification codes in the historical sensor data, wherein the historical sensor data is collected earlier than the current sensor data; In response to a quantity difference between the current identification code quantity and the historical identification code quantity being greater than a preset quantity difference threshold, the current sensor data is determined as the abnormal data.

8. The method according to claim 5, characterized in that After determining the identification code recognized by the camera on one side as the sensor data according to the number of identification codes recognized by the camera on each side, the method further includes: Obtaining an identification distance between the robot and a current identification code in the sensor data; Whether the sensor data is abnormal is determined according to the recognition distance, the number of identification codes currently recognized, and a preset number threshold; the recognition distance is positively correlated with the preset number threshold.

9. The method according to claim 5, characterized in that The robot further includes a left photoelectric sensor and a right photoelectric sensor respectively arranged according to the driving direction of the robot. After the response that the area type is a camera recognition area, the method further includes: Acquire the left photoelectric distance recognized by the left photoelectric sensor and the right photoelectric distance recognized by the right photoelectric sensor, and acquire the left identification distance between the robot and the current left identification code and the right identification distance between the robot and the current right identification code through the left camera and the right camera; In response to a distance difference between the left photoelectric distance and the left identification distance being greater than a preset distance difference threshold, determining the current left identification code as the abnormal data; Alternatively, if the distance difference between the right photoelectric distance and the right identification distance is greater than a preset distance difference threshold, the current right identification code is determined as the abnormal data.

10. A computer-readable storage medium having program instructions stored thereon, characterized in that: When the program instructions are executed by a processor, the method according to any one of claims 1 to 9 is implemented.

Images