Method for determining positioning precision of humanoid robot and related equipment

Abstract

The embodiment of the invention provides a method for determining the positioning precision of a humanoid robot and related equipment, and belongs to the technical field of robot control. According to the method, by obtaining multi-source positioning data from the humanoid robot and adjusting the weight coefficient of the multi-source positioning data, the multi-source positioning data can be combined with the original error value of the multi-source positioning data, and finally the positioning precision of the humanoid robot is determined. By adopting the method for determining the positioning precision, the weight coefficient of the multi-source positioning data can be flexibly adjusted according to different working environments of the humanoid robot, the accuracy of the positioning precision can be improved, and the working reliability of the humanoid robot under complex working conditions is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot control, and in particular to a positioning accuracy determination method for a humanoid robot and related equipment. BACKGROUND

[0002] In the working process of a humanoid robot, real-time positioning of the humanoid robot is needed to provide support for the work of the humanoid robot. Whenever the humanoid robot works for a period of time, the positioning accuracy of the humanoid robot needs to be detected to calibrate the positioning error of the humanoid robot in a timely manner. However, in actual working conditions, the working environment of the humanoid robot is relatively complex, and the positioning accuracy detection method of the related technology cannot adapt to environmental changes, and the accuracy of the positioning accuracy is relatively low, which cannot achieve the purpose of rapid, simple, and accurate detection. SUMMARY

[0003] The main purpose of the embodiments of the present application is to provide a positioning accuracy determination method for a humanoid robot and related equipment, which can quickly, simply, and accurately realize positioning accuracy detection of a humanoid robot according to multi-source data of the humanoid robot.

[0004] To achieve the above purpose, one aspect of an embodiment of the present application provides a positioning accuracy determination method for a humanoid robot, which comprises: obtaining multi-source positioning data collected by the humanoid robot, wherein the multi-source positioning data comprises radar positioning data, visual sensor positioning data, and robot pose positioning data; determining a weight coefficient of each of the multi-source positioning data of the humanoid robot according to a data reliability of the multi-source positioning data; determining the positioning accuracy of the humanoid robot in combination with the weight coefficient and an original error value of the multi-source positioning data.

[0005] In some embodiments, the determining of the weight coefficient of each of the multi-source positioning data of the humanoid robot according to the data reliability of the multi-source positioning data comprises: determining a working scene type of the humanoid robot according to the multi-source positioning data, wherein the working scene type comprises an industrial scene and a home scene; determining a working state of the humanoid robot according to the multi-source positioning data, wherein the working state comprises a moving state and a working state; determining the weight coefficient of each of the multi-source positioning data when the humanoid robot is in the working scene type and / or in the working state.

[0006] In some embodiments, the determining of the weight coefficient of each of the multi-source positioning data of the humanoid robot comprises: determining a data error rate of each of the multi-source positioning data; determine data reliability of each of the multi-source positioning data according to the data error rate; determine a weight coefficient of each of the multi-source positioning data based on the data reliability.

[0007] In some embodiments, the data error rate includes a radar point cloud matching error rate, a visual matching error rate, and a pose solving error rate, and the data reliability includes radar data reliability, visual data reliability, and pose data reliability; and the determining the data reliability of each of the multi-source positioning data according to the data error rate comprises: determining radar data reliability of the radar positioning data according to the radar point cloud matching error rate; determining visual data reliability of the visual sensor positioning data according to the visual matching error rate; determining pose data reliability of the multi-source positioning data according to the pose solving error rate.

[0008] In some embodiments, the determining the positioning accuracy of the humanoid robot in combination with the weight coefficient and the original error value of the multi-source positioning data comprises: obtaining a scene correction value corresponding to the working scene type; determining the positioning accuracy of the humanoid robot in combination with the weight coefficient, the original error value of the multi-source positioning data, and the scene correction value.

[0009] In some embodiments, the determining the positioning accuracy of the humanoid robot in combination with the weight coefficient and the original error value of the multi-source positioning data comprises: obtaining a motion correction value corresponding to the working state of the humanoid robot; determining the positioning accuracy of the humanoid robot in combination with the weight coefficient, the original error value of the multi-source positioning data, and the motion correction value.

[0010] In some embodiments, the determining the working scene type of the humanoid robot according to the multi-source positioning data comprises: determining an occluded area of the humanoid robot according to the multi-source positioning data; determining the working scene type of the humanoid robot in combination with the occluded area.

[0011] In some embodiments, if the occluded area is greater than a first threshold, the determining the weight coefficient of each of the multi-source positioning data when the humanoid robot is in the working scene type and / or in the working state comprises: obtaining offline map data; adding the offline map data into the multi-source positioning data to generate emergency positioning data; determining a weight coefficient of each of the emergency positioning data of the humanoid robot as a preset weight value.

[0012] To achieve the above object, another aspect of the embodiment of the present application provides a positioning accuracy determination device of a humanoid robot, which comprises: an acquisition module configured to acquire multi-source positioning data collected by the humanoid robot, wherein the multi-source positioning data comprises radar positioning data, visual sensor positioning data and robot pose positioning data; a calculation module configured to determine a weight coefficient of each of the multi-source positioning data of the humanoid robot according to a data reliability of the multi-source positioning data; a determination module configured to determine a positioning accuracy of the humanoid robot in combination with the weight coefficient and an original error value of the multi-source positioning data.

[0013] To achieve the above object, another aspect of the embodiment of the present application provides a humanoid robot, which comprises at least one radar, at least one visual sensor, at least one pose sensor, and a controller connected to the radar, the visual sensor and the pose sensor respectively: The controller is configured to acquire multi-source positioning data collected by the humanoid robot, wherein the multi-source positioning data comprises radar positioning data from the radar, visual sensor positioning data from the visual sensor and robot pose positioning data from the pose sensor; determine a weight coefficient of each of the multi-source positioning data of the humanoid robot according to a data reliability of the multi-source positioning data; and determine a positioning accuracy of the humanoid robot in combination with the weight coefficient and an original error value of the multi-source positioning data.

[0014] To achieve the above object, another aspect of the embodiment of the present application provides an electronic device, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the above method when executing the computer program.

[0015] To achieve the above object, another aspect of the embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program implements the above method when executed by a processor.

[0016] To achieve the above object, another aspect of the embodiment of the present application provides a computer program product, which comprises a computer program, and the computer program implements the above method when executed by a processor.

[0017] The embodiments of the present application at least have the following beneficial effects: the present application provides a positioning accuracy determination method and device for a humanoid robot, an electronic device, a storage medium and a program product, the scheme obtains multi-source positioning data from the humanoid robot, adjusts the weight coefficient of the multi-source positioning data, makes the multi-source positioning data combine the original error value of the multi-source positioning data, and finally determines the positioning accuracy of the humanoid robot. The positioning accuracy determination method of the present application can flexibly adjust the weight coefficient of the multi-source positioning data according to the different working environment of the humanoid robot, can improve the accuracy of the positioning accuracy, and ensure the reliability of the humanoid robot working in complex working conditions. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 is a flowchart of the positioning accuracy determination method for a humanoid robot provided by the embodiments of the present application; Figure 2 is a flowchart of step S102 in Figure 1 Figure 3 is a flowchart of step S103 in Figure 1 Figure 4 is another flowchart of step S102 in Figure 1 Figure 5 is another flowchart of step S103 in Figure 1 Figure 6 is another flowchart of step S102 in Figure 1 Figure 7 is another flowchart of step S102 in Figure 1 Figure 8 is a structural schematic diagram of the positioning accuracy determination device for a humanoid robot provided by the embodiments of the present application; Figure 9 is a hardware structural schematic diagram of the electronic device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical scheme and advantages of the present application clearer, the following will further describe the present application in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application. When the following description relates to the drawings, the same numbers in different drawings represent the same or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application, but are only examples of devices and methods consistent with some aspects of the embodiments of the present application as described in the appended claims.​​​​​​

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description herein is for describing the embodiments of the application only and is not intended to be limiting of the application.

[0021] Before the embodiments of the present application are explained in detail, the terminology and some of the names involved in the embodiments of the present application are first explained, and the terminology and names involved in the embodiments of the present application are applicable to the following explanations.

[0022] 1) humanoid robot: a robot designed to imitate the appearance and behavior of humans.

[0023] 2) positioning accuracy: the closeness between the spatial entity position information (usually coordinates) and its true position.

[0024] 2) original error value: refers to the error value that may exist in each sensor itself when the humanoid robot performs positioning data acquisition, for example, the error of the point cloud generated when the radar scans around.

[0025] In the related art, when the positioning accuracy of the robot is determined, a fixed parameter accuracy determination method is generally used. However, compared with general robots, due to the complex structure and working environment of humanoid robots, the existing accuracy determination method is difficult to apply, which may cause the problem of inaccurate positioning accuracy of the humanoid robot.

[0026] Therefore, in the embodiments of the present application, a humanoid robot positioning accuracy determination method and related equipment are provided, which can combine the multiple positioning data that the humanoid robot can receive, adjust the weight coefficient of each positioning data, so that the positioning accuracy of the humanoid robot can adapt to different working scenes or working states, and ensure the reliability of the positioning accuracy.

[0027] The method for determining positioning accuracy of a humanoid robot and related equipment provided in the embodiments of the present application relate to the technical field of accuracy detection. The method for determining positioning accuracy of a humanoid robot provided in the embodiments of the present application can be applied to a terminal, can also be applied to a server, and can further be software running in the terminal or the server. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted terminal, and the like, but is not limited thereto; the server end can be configured as a stand-alone physical server, can also be configured as a server cluster or a distributed system formed by multiple physical servers, can further be configured as a cloud server providing basic cloud computing services such as cloud service, cloud database, cloud computing, cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN, and big data and artificial intelligence platform, and the server can also be a node server in a blockchain network; and the software can be an application for implementing the method for determining positioning accuracy of a humanoid robot, and the like, but is not limited to the above forms.

[0028] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment, in which tasks are performed by remote processing devices connected by a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0029] The present application relates to a humanoid robot, comprising at least one radar, at least one vision sensor, at least one pose sensor, and a controller connected with the radar, the vision sensor and the pose sensor respectively. The method for determining positioning accuracy of the present embodiment is executed in the controller.

[0030] Figure 1 is an optional flowchart of the method for determining positioning accuracy of a humanoid robot provided in the embodiments of the present application, Figure 1 The method in can include, but is not limited to, steps S101 to S103.

[0031] In step S101, multi-source positioning data collected by a humanoid robot is acquired, and the multi-source positioning data includes radar positioning data, vision sensor positioning data, and robot pose positioning data.

[0032] The humanoid robot is installed with at least one radar, at least one vision sensor and at least one pose sensor. The radar sensor is used to scan around the humanoid robot and generate point cloud data, the generated point cloud data is used as radar positioning data, the vision sensor is used to take images around the humanoid robot, and the pose sensor is used to determine the action of the humanoid robot. The humanoid robot can realize its own positioning based on the radar, vision sensor and pose sensor.

[0033] In step S102, the weight coefficients of the multi-source positioning data of the humanoid robot are determined according to the data reliability of the multi-source positioning data.

[0034] On the basis of determining the multi-source positioning data, the embodiment further realizes the weight coefficients of the multi-source positioning data, so that the proportion of each multi-source positioning data changes, and then adapts to different working scenes and working states of the humanoid robot.

[0035] In step S103, the positioning accuracy of the humanoid robot is determined by combining the weight coefficients and the original error values of the multi-source positioning data.

[0036] On the basis of determining the weight coefficients of the multi-source positioning data, the embodiment further combines the original error values of the sensors corresponding to the weight coefficients, and finally determines the positioning accuracy of the humanoid robot by weighted calculation of the weight coefficients and the original error values.

[0037] In some embodiments, the calculation method of the weight coefficients and the original error values can be: (1) P is the final positioning accuracy (unit: cm); i represents any one of the radar, the vision sensor and the pose sensor, L represents the radar positioning data, V represents the vision sensor positioning data, I represents the robot pose positioning data; P i is the original error value of each data source (radar positioning data P L , vision positioning data P V , robot pose positioning data IMUP I (unit: cm), W i is the weight coefficient of the multi-source positioning data.

[0038] The steps S101 to S103 shown in the embodiments of the present application, the controller of the humanoid robot can be connected to the radar, the visual sensor and the pose sensor respectively, and after obtaining the positioning data of each sensor, further realizes the adaptive modification of the weight coefficient of the multi-source positioning data, and combines the original error value for combined calculation, realizes the determination of the positioning accuracy. The design idea of the present embodiment is that when the error of the sensor is large, the weight of the positioning data corresponding to the sensor is adaptively reduced, and the influence of high error data on the final accuracy detection is reduced. The present embodiment can adaptively adjust the positioning accuracy of the humanoid robot in complex working scenes and working states, ensure the accuracy of positioning, make the humanoid robot adapt to complex working conditions, and improve the working reliability of the humanoid robot.

[0039] By adaptively adjusting the working mode of the sensor, the data processing energy consumption can be reduced. The average power consumption of the industrial scene sensor is reduced from 15W to 11.25W (reduced by 25%), and the average power consumption of the household scene sensor is reduced from 10W to 7W (reduced by 30%).

[0040] In some embodiments, the environment image can be obtained by the visual sensor, and the data in the environment image can be used as the visual sensor positioning data. The natural image and the artificial image can be displayed in the environment image respectively. The difference between the natural image and the artificial image mainly considers whether the object displayed in the image is artificially preset. For example, the image corresponding to the object such as the corner of furniture, the texture of the wall surface and the outline of the equipment displayed in the image can be divided into natural image data, and the image of the artificial marker such as the added two-dimensional code and the Augment Reality (AR) displayed in the image can be divided into artificial image data.

[0041] In some embodiments, if the artificial marker is displayed in the environment image, the artificial marker can be further combined with formula (1) to realize the calculation of the positioning accuracy. The formula of the artificial marker can be: (2) M represents the artificial marker, W M is the weight coefficient of the artificial marker, P M is the original error value of the artificial marker.

[0042] When calculating the weight coefficient of the multi-source positioning data, the artificial image data in the environment image can dynamically affect the final weight coefficient, so as to finally affect the positioning accuracy of the humanoid robot. For example, in the environment image including more artificial markers, the influence of the artificial image data on the weight coefficient is relatively strong, and in the environment image including fewer artificial markers, the influence of the artificial image data on the weight coefficient is relatively weak.

[0043] In some embodiments, as shown in Figure 2 Step S102, according to the data reliability of the multi-source positioning data, the weight coefficient of each multi-source positioning data of the humanoid robot is determined, including: Step S201, according to the multi-source positioning data, the working scene type of the humanoid robot is determined, and the working scene type includes an industrial scene and a home scene. Step S202, the weight coefficient of each multi-source positioning data of the humanoid robot in the working scene type is determined.

[0044] The technical scheme of the embodiment further illustrates the acquisition method of the multi-source positioning data. When the multi-source positioning data is acquired, the working scene type of the humanoid robot is determined, and the working scene type can include an industrial scene and a home scene. The identification of the working scene type can be realized by identifying the objects contained in the surrounding environment. For example, it can be determined to be an industrial scene by identifying shelves, industrial equipment and the like contained in the scene, and it can be determined to be a home scene by identifying household appliances, furniture and the like contained in the scene. In different working scenes, the humanoid robot can perform different specifications of positioning accuracy determination, and be used to guide the work of the humanoid robot. For example, in an industrial scene, the robot needs to realize ±2cm level precision material grabbing in a warehouse with dense shelves, and complete ±5cm level collaborative positioning in an open production line area; in a home scene, it needs to realize ±10cm level obstacle avoidance navigation and fixed point service in a living room and a corridor blocked by furniture.

[0045] Illustratively, the radar can be a laser radar, and the laser radar acquires environmental point cloud data and acquires an effective point cloud proportion, which is the ratio between unblocked point cloud data and total scanning point cloud data. The visual sensor can identify the surrounding occlusion and determine the area proportion of the occlusion. By identifying the effective point cloud proportion and the area proportion of the occlusion, the working scene type can be determined. When the area proportion of the occlusion is in the interval [0.1, 0.5] (pieces / m²) and the effective point cloud proportion is in the interval [60%, 90%], it can be determined to be an industrial scene; when the area proportion of the occlusion is in the interval [0.2, 0.6] (pieces / m²) and the effective point cloud proportion is in the interval [50%, 85%], it can be determined to be a home scene.

[0046] In some embodiments, as shown in Figure 7 Step S201, according to the multi-source positioning data, the working scene type of the humanoid robot is determined, including: Step S701, according to the multi-source positioning data, the area of the occlusion region around the humanoid robot is determined. Step S702, in combination with the area of the occlusion region, the working scene type of the humanoid robot is determined.

[0047] In some embodiments, in addition to determining the industrial scene and the home scene, the determined working scene type can also determine the occlusion density in combination with the radar positioning data, and determine the specific area in the industrial scene or the home scene. For example, in the industrial scene, the occlusions such as shelves, equipment, etc. can be identified and determined as open operation area, shelf dense area or equipment occlusion area; in the home scene, the occlusions such as sofa, wardrobe, etc. can be identified and determined as open living room, corridor passage or furniture occlusion area, etc.

[0048] For example, Table 1.1 lists a plurality of working scenes and corresponding determination methods.

[0049]

[0050] Table 1.1: A plurality of working scenes and corresponding determination methods The identification of the working scene type of the humanoid robot is realized by identifying the point cloud generated by the laser radar sensor scanning and the scene identifier. For example, the laser radar (horizontal field of view 360°, vertical field of view -15°~+15°, point cloud density 100 points / ㎡) scans a total of 10000 point clouds, 7200 unoccluded point clouds, effective point cloud proportion ; 20 shelves (occlusions) are identified, the robot perception range area is 50m², the occlusion density ; 8 two-dimensional code markers are identified, the marker distribution density . Compared with the threshold interval , the target scene is determined to be “industrial shelf dense area” The robot enters the bedroom of the family (furniture occlusion area), the laser radar (horizontal field of view 120°, vertical field of view 30°, point cloud density 80 points / ㎡) scans a total of 8000 point clouds, 3600 unoccluded point clouds, ; 15 furniture (bed, wardrobe, etc.) are identified, the perception range area is 20m², ; 1 AR marker is identified, , the target scene is determined to be “home furniture occlusion area”.

[0051] The technical solution of the present embodiment can acquire multi-source positioning data in combination with the working scene type of the humanoid robot, so that the multi-source positioning data is more targeted and can better adapt to different working environments.

[0052] In some embodiments, when acquiring multi-source positioning data for different working scene types, the positioning accuracy can also be further weighted and calculated for different scenes when performing positioning accuracy calculation. As shown in Figure 3 , in step S103, the positioning accuracy of the humanoid robot is determined in combination with the weight coefficient and the original error value of the multi-source positioning data, including: Step 301, obtaining a scene correction value corresponding to the working scene type; Step 302, combining the weight coefficient, the original error value of the multi-source positioning data and the scene correction value, determining the positioning accuracy of the humanoid robot.

[0053] For different working scenes, the technical scheme of the embodiment proposes a scene correction value corresponding to the working scene type when calculating the positioning accuracy. Considering the influence of different working scene types on the multi-source positioning data, the scene correction value proposed in the embodiment is introduced to correct the positioning accuracy value when calculating the positioning accuracy. Different regions may exist in different scenes. For example, in an industrial scene, it can include an open operation area, a dense shelf area or a device shielding area; in a home scene, it can include a home corridor area, a home open living room area, etc. Different scene correction values can be set for different regions in each working scene type. When the humanoid robot works in different working scenes, it can cover different regions in the working scene. When calculating the positioning accuracy, the scene correction value is introduced to reduce the influence of scene factors on the positioning accuracy.

[0054] When the scene correction value is introduced to calculate the positioning accuracy, the calculation method can be: (3) P is the final positioning accuracy, S is the scene correction factor, K S is the original error value, which is different for different working types S Different, for example, the industrial open area S =0.1, the industrial dense shelf area S =0.3, the industrial device shielding area S =0.6; the home open living room S =0.15, the home corridor passage S =0.4, the home furniture shielding area S =0.7; the shielding object distribution in the home scene is more irregular, resulting in a sensor data loss rate 10%-15% higher than that of the industrial regular shielding object, so the home full shielding area S The value is slightly higher than that of the industrial full shielding area; K S is the scene error reference value (fixed at 10 cm, representing the basic influence of the scene on positioning).

[0055] In some embodiments, when multi-source positioning data is acquired for different working scene types, the positioning accuracy can be further calculated by weighting for different working states of the robot when calculating the positioning accuracy. The working state of the robot can include the moving state and the working state of the robot, wherein the moving state can include the position and the moving speed of the robot, and the working state can be used to represent the type of work performed by the robot or the action such as grabbing performed by the robot control arm. When acquiring multi-source positioning data, the humanoid robot can have certain influence on the value of the positioning data due to the position or pose thereof, and therefore, when acquiring multi-source positioning data, the working state of the humanoid robot needs to be determined, such as Figure 4 As shown in FIG. 10, in step S102, the weight coefficient of each multi-source positioning data of the humanoid robot is determined according to the data reliability of the multi-source positioning data, including: In step S401, the working state of the humanoid robot is determined according to the multi-source positioning data, and the working state includes the moving state and the working state.

[0056] The robot speed is acquired by the pose sensor (acceleration measurement range ±18g, angular velocity measurement range ±250dps, and update frequency 100Hz) and the vision sensor (based on ORB-SLAM3 algorithm) at 0.6m / s, the moving state and the working state are “fast turning + obstacle avoidance”, and the working state is determined to be “high-speed working state”.

[0057] In step S402, the weight coefficient of each multi-source positioning data of the humanoid robot in the working state is determined.

[0058] When calculating the positioning accuracy, the moving state and the working state of the humanoid robot can be compensated when calculating the positioning accuracy based on the extracted multi-source positioning data. As shown in Figure 5 In step S103, the positioning accuracy of the humanoid robot is determined by combining the weight coefficient and the original error value of the multi-source positioning data, including: In step S501, a motion correction value corresponding to the working state of the humanoid robot is acquired. In step S502, the positioning accuracy of the humanoid robot is determined by combining the weight coefficient, the original error value of the multi-source positioning data, and the motion correction value.

[0059] To address different working states when calculating positioning accuracy, this embodiment proposes motion correction values ​​corresponding to each working state. Considering the impact of different working states on multi-source positioning data, the motion correction values ​​proposed in this embodiment are used to correct the positioning accuracy value during the calculation. Different working states correspond to different movement and operation states. For example, a humanoid robot may be in a linear movement state while simultaneously in a slow turning operation state; a humanoid robot may also be in a rapid movement state while simultaneously in a continuous grasping operation state. Introducing motion correction values ​​during positioning accuracy calculation reduces the impact of working state factors on positioning accuracy.

[0060] For example, as shown in Table 1.2 below, under different working conditions, the data reliability of different multi-source positioning data varies.

[0061]

[0062] Table 1.2 Multi-source positioning data under different working conditions Therefore, before determining the multi-source localization data, the working state of the humanoid robot can be further determined. When introducing scene correction values ​​to calculate localization accuracy, the calculation method can be as follows: (4) M and K M The product value is used as the motion correction value; M is the motion correction factor, where M=0.05 for stationary state, M=0.2 for low-speed movement state, and M=0.5 for high-speed operation state; K M The baseline value for motion error is 8cm (fixed to represent the basic influence of motion state on positioning).

[0063] The values ​​of scene correction factor S and motion correction factor M are obtained by collecting no less than 50 sets of robot positioning error samples in the corresponding work scene type / work state, and fitting them through correlation analysis between the error samples and scene / motion parameters, with a fitting error ≤5%.

[0064] In some embodiments, multi-source positioning data may be affected by both the work scene type and the work status. When calculating positioning accuracy, scene correction values ​​and motion correction values ​​can be considered simultaneously. The calculation method is as follows: (5) In some embodiments, such as Figure 6 As shown, step S102 involves determining the weighting coefficients of each multi-source localization data point for the humanoid robot based on the data reliability of the multi-source localization data, including: Step S601, determine the data error rate of each multi-source positioning data; Step S602, according to the data error rate, determine the data reliability of each multi-source positioning data; Step S603, based on the data reliability, determine the weight coefficient of each multi-source positioning data.

[0065] The data error rate is the error rate of the sensor itself when the humanoid robot carries the sensor to obtain each data. When calculating the weight coefficient, the data error rate of each sensor itself is combined to realize the calculation of data reliability and further realize the determination of weight coefficient.

[0066] Data reliability R i The calculation method can be: (6) i represents any one of laser radar (L), visual sensor (V), pose sensor (I) and marker (M), represents the error rate, indicates the data update frequency.

[0067] Data reliability of laser radar R L The calculation method can be: (7) is the laser radar point cloud matching error rate (unit: %), which is calculated by ICP algorithm to calculate the average deviation of matching point pairs and the ratio of scanning distance, that is, ; is the point cloud data update frequency (unit: Hz).

[0068] Data reliability of visual sensor R V The calculation method can be: (8) is the visual feature matching error rate (unit: %), that is, the proportion of error matching points of ORB feature matching; is the image data update frequency (unit: Hz); is the environmental marker distribution density (unit: pieces / m²).

[0069] Data reliability of pose sensor R I The calculation method can be: (9) is the pose sensor pose solving error rate (unit: %); is the pose sensor data update frequency (unit: Hz); is the robot motion acceleration (unit: m / s²), the value range is [0, 1], when the actual acceleration of the robot exceeds 1 m / s², it is counted as 1 m / s²; the greater the acceleration, the lower the reliability of the pose sensor. Data reliability of the marker R M The calculation method can be: (10) is the marker coordinate recognition error rate (unit: %); is the marker recognition success rate (unit: %).

[0070] After the reliability calculation is implemented respectively, the further weight coefficient calculation method can be:

[0071] (11) W i is the weight coefficient of multi-source positioning data.

[0072] In some embodiments, the collected positioning data further includes a pre-stored environment map, which is constructed through robot simultaneous localization and mapping (SLAM) and supports online incremental updating.

[0073] In some embodiments, when the robot enters the warehouse shelf display area, the laser radar (horizontal field of view 360°, vertical field of view -15°~+15°, point cloud density 100 points / ㎡) scans a total of 10,000 point clouds, 7,200 unobstructed point clouds, and the effective point cloud proportion R cloud =72%; 20 shelves (obstructions) are identified, the robot perception range area is 50 m², the obstruction density D obs =20 / 50=0.4 / m²; 8 two-dimensional code markers are identified, the marker distribution density D mark =8 / 50=0.16 / m². By comparing the threshold interval, it is determined that the target scene is an “industrial shelf dense area” (0.1 D obs ≤0.5, 60%≤ R cloud <90%).

[0074] The robot speed 0.3 m / s is obtained by the position sensor (acceleration measurement range ± 16g, angular velocity measurement range ± 2000 dps, update frequency 50 Hz) and the wheel encoder, the action type is "straight line movement + intermittent grabbing", and the motion state is "low speed movement state".

[0075] Synchronous acquisition of sensor data, laser radar point cloud matching error rate (through ICP algorithm calculation, average deviation of matching points 0.06 m, scanning distance 3 m, , update frequency ; visual feature matching error rate (ORB feature error matching point ratio 3%), update frequency ; IMU pose solving error rate , update frequency , acceleration (≤1 m / s², according to the actual value) ; marker coordinate recognition error rate , recognition success rate .

[0076] Reliability calculation:

[0077]

[0078]

[0079] (12) Weight coefficient calculation:

[0080]

[0081] (13) Precision calculation: known original error of each sensor ; scene correction factor S = 0.3 (industrial shelf dense area), motion correction factor M = 0.2 (low speed movement state); substitute into the formula to obtain:

[0082] That is, the positioning accuracy of the robot in this scene is about 8.4 cm, which meets the accuracy requirement of warehouse material handling ± 10 cm.

[0083] In some embodiments, in some special environments, for example, the shielding area around the humanoid robot, that is, when the shielding area is greater than the first threshold, such as Figure 7As shown, in step S102, the weight coefficients of each multi-source positioning data of the humanoid robot are determined according to the data reliability of the multi-source positioning data, and the method can further include: In step S703, offline map data is acquired. In step S704, the offline map data is added to the multi-source positioning data to generate emergency positioning data. In step S705, the weight coefficient of each emergency positioning data of the humanoid robot is determined as a preset weight value.

[0084] When the target scene is a full occlusion area (industrial equipment occlusion area / household furniture occlusion area) and the laser radar reliability coefficient ( The value range is [0, 1]) and the visual sensor reliability coefficient is within 2 seconds, the emergency positioning mode is started: the offline map data is acquired, only the matching data of the pose sensor and the pre-stored offline map data is retained, the dynamic weight coefficient is adjusted to ( The map matching weight of the pre-stored map), and the positioning accuracy formula is updated to , wherein is the map matching error (unit: cm).

[0085] The emergency positioning mode covers the full occlusion scene and solves the "positioning blind state" problem. The emergency positioning accuracy of the industrial equipment occlusion area can reach ±12 cm, and the household furniture occlusion area can reach ±18 cm, which respectively meet the basic requirements of industrial equipment inspection and household obstacle avoidance navigation.

[0086] Please refer to Figure 8 The embodiment of the present application also provides a positioning accuracy determination device of a humanoid robot, which can implement the above method. The device comprises: An acquisition module 801 is configured to acquire multi-source positioning data collected by the humanoid robot, wherein the multi-source positioning data comprises radar positioning data, visual sensor positioning data, robot pose positioning data and map positioning data. A calculation module 802 is configured to determine weight coefficients of each multi-source positioning data of the humanoid robot according to data reliability of the multi-source positioning data. A determination module 803 is configured to determine positioning accuracy of the humanoid robot by combining the weight coefficients and original error values of the multi-source positioning data.

[0087] In some embodiments, the calculation module 802 is configured to: determine a working scene type of the humanoid robot according to the multi-source positioning data, wherein the working scene type comprises an industrial scene and a household scene. determine a working state of the humanoid robot according to the multi-source positioning data, the working state comprising a moving state and a working state; determine a weight coefficient of each of the multi-source positioning data when the humanoid robot is in the working scene type and / or in the working state.

[0088] In some embodiments, the computing module 802 is configured to: determine a data error rate of each of the multi-source positioning data; determine a data reliability of each of the multi-source positioning data according to the data error rate; determine a weight coefficient of each of the multi-source positioning data based on the data reliability.

[0089] In some embodiments, the data error rate comprises a radar point cloud matching error rate, a visual matching error rate and a pose solving error rate, and the data reliability comprises a radar data reliability, a visual data reliability and a pose data reliability; the computing module 802 is configured to: determine a radar data reliability of the radar positioning data according to the radar point cloud matching error rate; determine a visual data reliability of the visual sensor positioning data according to the visual matching error rate; determine a pose data reliability of the multi-source positioning data according to the pose solving error rate.

[0090] In some embodiments, the determining module 803 is configured to: obtain a scene correction value corresponding to the working scene type; determine a positioning accuracy of the humanoid robot in combination with the weight coefficient, an original error value of the multi-source positioning data and the scene correction value.

[0091] In some embodiments, the determining module 803 is configured to: obtain a motion correction value corresponding to the working state of the humanoid robot; determine a positioning accuracy of the humanoid robot in combination with the weight coefficient, an original error value of the multi-source positioning data and the motion correction value.

[0092] In some embodiments, the computing module 802 is configured to: determine an occluded area of the humanoid robot according to the multi-source positioning data; determine a working scene type of the humanoid robot in combination with the occluded area.

[0093] In some embodiments, if the occluded area is greater than a first threshold, the computing module 802 is configured to: obtain offline map data; adding the offline map data into the multi-source positioning data to generate emergency positioning data; determining a weight coefficient of each of the emergency positioning data of the humanoid robot as a preset weight value.

[0094] It can be understood that the contents in the above method embodiments are all applicable to the device embodiments, the device embodiments specifically implement the functions of the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0095] The embodiment of the present application further provides an electronic device, which comprises a memory and a processor, the memory stores a computer program, and the processor implements the above method when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a vehicle-mounted computer, etc.

[0096] It can be understood that the contents in the above method embodiments are all applicable to the device embodiments, the device embodiments specifically implement the functions of the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0097] Please refer to Figure 9 , Figure 9 The hardware structure of the electronic device of another embodiment is illustrated, which comprises: The processor 901 can be implemented in the form of a general CPU (Central Processing Unit), a microprocessor, an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits, etc., for executing related programs to implement the technical solutions provided by the embodiments of the present application; The memory 902 can be implemented in the form of a ROM (Read Only Memory), a static storage device, a dynamic storage device, or a RAM (Random Access Memory), etc. The memory 902 can store an operating system and other application programs, and when the technical solutions provided by the embodiments of the present application are implemented by software or firmware, the related program codes are saved in the memory 902 and called and executed by the processor 901 to implement the above method of the embodiments of the present application; The input / output interface 903 is used to realize information input and output; The communication interface 904 is used to realize the communication interaction between the device and other devices, which can realize communication through a wired manner (for example, a USB, a network cable, etc.) or a wireless manner (for example, a mobile network, WIFI, Bluetooth, etc.). a bus 905 that transmits information between various components (e.g., the processor 901, the memory 902, the input / output interface 903, and the communication interface 904) of the device; The processor 901, the memory 902, the input / output interface 903, and the communication interface 904 are communicatively connected to each other within the device through the bus 905.

[0098] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the method.

[0099] It can be understood that the contents in the above method embodiments are all applicable to the storage medium embodiment, the storage medium embodiment specifically implements the functions of the above method embodiments, and achieves the same beneficial effects as the above method embodiments.

[0100] The embodiment of the present application further provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the method.

[0101] It can be understood that the contents in the above method embodiments are all applicable to the program product embodiment, the program product embodiment specifically implements the functions of the above method embodiments, and achieves the same beneficial effects as the above method embodiments.

[0102] The memory is a non-transitory computer readable storage medium, which can be used to store a non-transitory software program and a non-transitory computer executable program. In addition, the memory can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory remotely arranged relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0103] The embodiment of the present application provides a positioning accuracy determination method and device of a humanoid robot, an electronic device, a storage medium and a program product. The positioning accuracy determination method of the humanoid robot comprises the following steps: obtaining multi-source positioning data from the humanoid robot; adjusting a weight coefficient of the multi-source positioning data, so that the multi-source positioning data can combine an original error value of the multi-source positioning data; and finally determining the positioning accuracy of the humanoid robot. The positioning accuracy determination scheme of the present application can flexibly adjust the weight coefficient of the multi-source positioning data according to different working environments of the humanoid robot, can improve the accuracy of the positioning accuracy, and can ensure the reliability of the humanoid robot working in complex working conditions.

[0104] The embodiments described in the specification are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0105] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and can include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0106] The device embodiments described above are merely illustrative, and the units described as separate components can or can not be physically separated, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0107] Those skilled in the art can understand that all or some of the steps in the above disclosed method, the functional modules / units in the system and the device can be implemented as software, firmware, hardware and their appropriate combinations.

[0108] The terms "first", "second", "third", "fourth" and the like (if any) in the specification of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0109] It should be understood that, in the application, "at least one" refers to one or more, and "multiple" refers to two or more. "And / or" is used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, "A and / or B" can represent three cases of only A, only B and A and B existing at the same time, wherein A and B can be singular or plural. The character " / " generally represents an "or" relationship between the front and rear associated objects. "At least one of the following" or the like means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b or c can represent a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b and c can be single or multiple.

[0110] In several embodiments provided in the application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the above units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed mutual units can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0111] The units described above as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on multiple network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0112] In addition, each functional unit in each embodiment of the application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0113] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in other words, the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program storage media.

[0114] The preferred embodiments of the embodiments of the present application are described above with reference to the accompanying drawings, and are not limited to the scope of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the embodiments of the present application.

Claims

1. A method for determining the positioning accuracy of a humanoid robot, characterized in that, include: Acquire multi-source localization data collected by the humanoid robot, including radar localization data, visual sensor localization data, and robot pose localization data; Based on the data reliability of the multi-source positioning data, the weighting coefficients of each of the multi-source positioning data of the humanoid robot are determined; The positioning accuracy of the humanoid robot is determined by combining the weighting coefficients and the original error values ​​of the multi-source positioning data.

2. The method according to claim 1, characterized in that, The step of determining the weighting coefficients of each of the multi-source positioning data for the humanoid robot based on the data reliability of the multi-source positioning data includes: The working scenario type of the humanoid robot is determined based on the multi-source positioning data, and the working scenario type includes industrial scenario and home scenario; The working state of the humanoid robot is determined based on the multi-source positioning data, and the working state includes a movement state and an operation state. Determine the weighting coefficients of the multi-source localization data for each of the humanoid robot in the work scenario type and / or the work state.

3. The method according to claim 1 or 2, characterized in that, The step of determining the weighting coefficient of each of the multi-source positioning data for the humanoid robot based on the data reliability value of the multi-source positioning data includes: Determine the data error rate of each of the multi-source positioning data; The data reliability of each of the multi-source positioning data is determined based on the data error rate. Based on the data reliability, the weighting coefficients of each of the multi-source positioning data are determined.

4. The method according to claim 3, characterized in that, The data error rate includes radar point cloud matching error rate, visual matching error rate, and pose calculation error rate; the data reliability includes radar data reliability, visual data reliability, and pose data reliability; determining the data reliability of each of the multi-source positioning data based on the data error rate includes: The radar data reliability of the radar positioning data is determined based on the radar point cloud matching error rate. The visual data reliability of the visual sensor positioning data is determined based on the visual matching error rate. The reliability of the pose data of the multi-source positioning data is determined based on the pose calculation error rate.

5. The method according to claim 2, characterized in that, Determining the positioning accuracy of the humanoid robot by combining the weighting coefficients and the original error values ​​of the multi-source positioning data includes: Obtain the scene correction value corresponding to the work scene type; The positioning accuracy of the humanoid robot is determined by combining the weighting coefficients, the original error values ​​of the multi-source positioning data, and the scene correction values.

6. The method according to claim 2, characterized in that, Determining the positioning accuracy of the humanoid robot by combining the weighting coefficients and the original error values ​​of the multi-source positioning data includes: Obtain the motion correction value corresponding to the working state of the humanoid robot; The positioning accuracy of the humanoid robot is determined by combining the weighting coefficients, the original error values ​​of the multi-source positioning data, and the motion correction values.

7. The method according to claim 2, characterized in that, Determining the working scenario type of the humanoid robot based on the multi-source positioning data includes: The area of ​​the obstructed region around the humanoid robot is determined based on the multi-source positioning data. Based on the area of ​​the obstructed region, the working scenario type of the humanoid robot is determined.

8. The method according to claim 7, characterized in that, If the area of ​​the occluded region is greater than a first threshold, determining the weighting coefficients of each of the multi-source localization data for the humanoid robot in the work scenario type and / or the work state includes: Obtain offline map data; Add the offline map data to the multi-source positioning data to generate emergency positioning data; The weighting coefficients of each of the emergency positioning data of the humanoid robot are determined to be preset weight values.

9. A humanoid robot, characterized in that, It includes at least one radar, at least one vision sensor, at least one pose sensor, and a controller connected to the radar, the vision sensor, and the pose sensor respectively. The controller is used to acquire multi-source positioning data collected by the humanoid robot. The multi-source positioning data includes radar positioning data from the radar, visual sensor positioning data from the visual sensor, and robot pose positioning data from the pose sensor. Based on the data reliability of the multi-source positioning data, the controller determines the weighting coefficient of each multi-source positioning data of the humanoid robot. Combining the weighting coefficients and the original error values ​​of the multi-source positioning data, the controller determines the positioning accuracy of the humanoid robot.

10. A device for determining the positioning accuracy of a humanoid robot, characterized in that, The device includes: The acquisition module is used to acquire multi-source positioning data collected by the humanoid robot, including radar positioning data, visual sensor positioning data and robot pose positioning data. The calculation module is used to determine the weighting coefficient of each of the multi-source positioning data of the humanoid robot based on the data reliability of the multi-source positioning data; The determination module is used to determine the positioning accuracy of the humanoid robot by combining the weighting coefficients and the original error values ​​of the multi-source positioning data.

Images