Data processing method and related system

Abstract

The embodiment of the invention provides a data processing method and a related system. The data processing method comprises the following steps: receiving data transmitted by the robot through the local network, wherein the data is acquired by a sensor; and according to the sensor distribution information corresponding to the robot, calling a corresponding service instance to perform frame alignment on the data to obtain a data processing result. The data processing method and the related system can reduce the calculation pressure of the robot.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of robotics, and more particularly to a data processing method and related systems. Background Art

[0002] A robot is a machine that can perform a series of tasks autonomously or under external control. It is composed of hardware, software, and sensors, and can interact with the environment through perception, execution, decision-making, etc. to perform specific tasks. In order to be able to perform a variety of tasks and adapt to different environments. Robots are usually equipped with a variety of data acquisition units, such as RGB cameras / depth cameras / lidar for collecting image data, inertial measurement units for collecting posture data, and force sensors for measuring the force and torque applied to the robot.

[0003] The data collected by the robot through sensors is usually large in volume, which can easily cause high computing pressure on the robot. Summary of the invention

[0004] In view of this, multiple embodiments of the present application are dedicated to providing a data processing method and related system, whereby the robot can utilize the local network to transmit data to a private cloud device for processing, thereby reducing the computing pressure of the robot.

[0005] One embodiment of the present application provides a data processing method, which is applied to a private cloud device, wherein the private cloud device and a robot transmit data via a local network, and the method comprises: receiving data transmitted by the robot via the local network, wherein the data is collected by a sensor; and based on the sensor distribution information corresponding to the robot, calling a corresponding service instance to perform frame alignment on the data to obtain a data processing result.

[0006] Optionally, the method also includes: the data is obtained by the robot compressing original data based on a compression rule, and the compression rule is used to limit the compression method corresponding to different data types.

[0007] Optionally, the data is obtained by the robot verifying data compression results based on verification rules, and the verification rules are used to limit at least one of the following verification dimensions: data integrity, data quality, system abnormality, interaction environment abnormality, and task execution abnormality.

[0008] Optionally, according to the sensor distribution information corresponding to the robot, calling the corresponding service instance to perform frame alignment on the data to obtain the step of data processing results, including: according to the sensor distribution information corresponding to the robot, calling the corresponding service instance and allocating corresponding computing resources to the service instance, so that the service instance determines the data list corresponding to each data acquisition unit in the robot based on the data, and checks the timestamp consistency of each group of data frames in turn through the created current frame group to obtain the data processing results; wherein, the data frames of any group of data frames are taken from each data list respectively.

[0009] Optionally, the private cloud device includes a data processing module and an object storage module; the step of receiving data transmitted by the robot through the local network, wherein the data is collected by sensors, includes: upon receiving a token request sent by the robot, sending a token for identifying a data collection cycle to the robot through the data processing module, so that the robot collects raw data based on the token, compresses and verifies the raw data into the data, and then uploads it to the object storage module, and returns the upload status of the data to the data processing module; upon receiving the upload status, triggering the data processing module to pull the data from the object storage module.

[0010] Optionally, the method also includes: labeling the data processing results according to labeling rules to obtain labeled data; wherein the labeling rules are used to indicate at least one data labeling dimension; uploading the labeled data to a public cloud device so that the public cloud device performs a model training task based on the labeled data; wherein the public cloud device and the private cloud device transmit data through an external network.

[0011] Optionally, before the step of labeling the data processing result according to labeling rules to obtain the labeled data, the method also includes: displaying the data processing result on a user review interface; and updating the data processing result when an adjustment operation on the data processing result is detected.

[0012] One embodiment of the present application provides a data processing device, which includes: a data receiving unit, used to receive data transmitted by the robot through the local network, the data being collected by sensors; a data processing unit, used to call a corresponding service instance to perform frame alignment on the data according to sensor distribution information corresponding to the robot, so as to obtain a data processing result.

[0013] One embodiment of the present application provides a data processing system, which includes a robot and a private cloud device, and the robot and the private cloud device transmit data through a local network, wherein: the robot is used to collect data through sensors and transmit the data to the private cloud device through the local network; the private cloud device is used to call the corresponding service instance to perform frame alignment on the data according to the sensor distribution information corresponding to the robot, so as to obtain a data processing result.

[0014] Optionally, the data processing system also includes a public cloud device, wherein: the public cloud device is used to receive the data processing results sent by the private cloud device.

[0015] One embodiment of the present application provides a computer device, which includes a memory and a processor, wherein the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor to implement the data processing method as described above.

[0016] One embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores at least one computer program, and when the at least one computer program is executed by a processor, the data processing method as described above can be implemented.

[0017] One embodiment of the present application provides a computer program product, which is used to implement the data processing method as described above.

[0018] In multiple implementations provided in the present application, data transmitted by the robot through a local network can be received, and the corresponding service instance can be called to perform frame alignment on the data based on the sensor distribution information corresponding to the robot to obtain data processing results, thereby reducing the computing pressure of the robot. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic diagram of a robot scheduling system provided for one embodiment of the present application.

[0020] Figure 2 A flowchart of a data processing method provided for one embodiment of the present application.

[0021] Figure 3 A schematic diagram of a module of a data processing device provided for one embodiment of the present application.

[0022] Figure 4 A schematic diagram of a computer device provided for one embodiment of the present application. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0024] In the description of the embodiments of the present application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0025] With the development of computer vision, artificial intelligence, sensor technology and other technologies, the application fields of robots are becoming more and more diverse. For example, robots can be applied to embodied intelligence, industrial automation, medical and health care, agricultural automation, intelligent services, detection, entertainment, and commercial data analysis. As the application fields of robots gradually expand, the complexity of robot hardware design gradually increases to enable them to adapt to more complex and changing environments.

[0026] In the related art, the robot may be provided with one or more of the following sensors: visual sensors, distance and proximity sensors, force and torque sensors, inertial measurement sensors, temperature, humidity and environment sensors, position and positioning sensors, sound and voice sensors, tactile and contact sensors, visual and optical sensors, biosensors, and air pressure sensors. The data collected by the sensors can be processed to generate further robot scheduling tasks. For example, the data collected by the distance and proximity sensors can be used to analyze the distance between the robot and obstacles, and the robot obstacle avoidance task can be generated based on the distance.

[0027] It is understandable that during the operation of the robot, the above-mentioned sensors can continuously collect data according to the requirements of the work tasks, and the amount of data collected is relatively large. Due to the size and specifications of the robot, the computing power of the robot is usually not too high. If the robot processes massive amounts of data, it is easy to have insufficient computing power, which will lead to reduced data processing efficiency.

[0028] Therefore, it is necessary to provide a data processing method to receive data transmitted by the robot through the local network, and to call the corresponding service instance to perform frame alignment on the data according to the sensor distribution information corresponding to the robot to obtain the data processing result, so as to reduce the computing pressure of the robot and reduce bandwidth consumption, delay risk, packet loss risk, and privacy leakage risk.

[0029] The embodiment of the present application provides a data processing system. The data processing method can be applied to the data processing system, the data processing system includes a robot and a private cloud device, the robot and the private cloud device perform data transmission through a local network, wherein: the robot is used to collect data through sensors and transmit the data to the private cloud device through the local network; the private cloud device is used to call the corresponding service instance to perform frame alignment on the data according to the sensor distribution information corresponding to the robot, and obtain the data processing result.

[0030] Optionally, the data processing system also includes a public cloud device, wherein: the public cloud device is used to receive the data processing results sent by the private cloud device.

[0031] Among them, private cloud devices are specially designed and deployed cloud computing infrastructures. Private cloud devices are used to provide private cloud services, such as data alignment services, to robots through local networks. Private cloud devices include servers for providing computing power, storage systems for storing various types of data, network devices for ensuring data transmission and data access control, etc. Robots are mechanical devices used to assist in the automation of specified fields. The robots are equipped with sensors for performing specific tasks in the instruction field. The types and quantities of sensors are not limited in the embodiments of this application. Public cloud devices are cloud computing services provided by third parties. Public cloud devices are used to provide public cloud services to robots through external networks.

[0032] Since the robot and the private cloud device are deployed in the same physical location (such as in the local area network of an enterprise), the private cloud device is the proximal device of the robot, and the network transmission between the two depends on the local network, that is, the local area network, which has a short communication distance and low latency. The public cloud device is deployed at the far end, and the data transmission between the robot and the public cloud device needs to pass through switches, routers, data centers and other devices. In other words, the data transmission process needs to pass through more transit nodes than the previous data transmission process. Therefore, the data transmission between the robot and the public cloud device obviously has the problem of long communication distance and high latency. In addition, the robot will also need to bear the risks of data packet loss and network jitter when transmitting data to the public cloud device. Based on this, in the data processing system, the private cloud device provides the robot with a more urgent data processing service for the robot, and the public cloud device further processes the data processing results of the private cloud device. Since the data processing results are often used for tasks such as model training tasks, these tasks are not urgent tasks for robots, so they can be completed using public cloud devices.

[0033] Specifically, see Figure 1. In the data processing system, the robot can perform data collection through the sensors installed on it to obtain data collection results. Furthermore, the robot can compress the data collection results to reduce the pressure of data transmission and improve the efficiency of data transmission. Furthermore, the robot can verify the data of the compression results to ensure the integrity and compliance of the transmitted data. Furthermore, the robot can upload the verified data to the private cloud device through the local network to trigger the private cloud device to process the data, which can avoid consuming too much robot computing power.

[0034] The private cloud device can perform data verification on the data to ensure the integrity of the data. Furthermore, the private cloud device can perform parsing control operations, specifically, according to the sensor distribution information corresponding to the robot, call the corresponding service instance and allocate computing resources suitable for the service instance. Furthermore, the private cloud device can perform frame alignment operations on the data through the service instance to obtain data processing results. Furthermore, the private cloud device can perform visualization processing on the data processing results, and the visualization processing can be performed according to a predetermined display rule. The visualization results obtained by the visualization processing can be used for display for data review. Furthermore, when the private cloud device displays the visualization results and detects the data review operation for the visualization results, it determines the data review results. The data review operation includes but is not limited to the deletion, modification, and confirmation of the visualization results. Furthermore, the private cloud device can perform data pre-annotation on the visualization results based on the annotation rules used to limit the annotation dimensions (such as whether it is sensitive data). Furthermore, the private cloud device can also detect data modification operations on the pre-annotated results, and the data modification operations are used to modify, delete, and confirm the pre-annotated labels. Furthermore, the private cloud device can upload the data labeling results to the database of the public cloud device through the external network, so that the public cloud device can perform model training tasks based on the labeled data that are not very urgent but require high computing power.

[0035] See also Figure 2 . An embodiment of the present application provides a data processing method. The data processing method can be applied to a private cloud device, and the private cloud device and the robot transmit data through a local network. It can also be applied to a data processing device, and the data processing device can be an electronic device with certain computing power and network access capabilities. Of course, the data processing device can also refer to a software program running in an electronic device. In this embodiment, the data processing method may include the following steps.

[0036] Step S110: receiving data transmitted by the robot through the local network, wherein the data is collected by sensors.

[0037] Step S120: According to the sensor distribution information corresponding to the robot, the corresponding service instance is called to perform frame alignment on the data to obtain a data processing result.

[0038] In this embodiment, in order to process robot data more efficiently and save bandwidth, the robot can efficiently upload data to the private cloud device through the local network to trigger the private cloud device to process the data.

[0039] Among them, the robot can be any of the following: a manufacturing robot used in industrial production lines or manufacturing fields, such as a welding robot, a handling robot, etc.; a service robot used to provide services to humans, such as a sweeping robot, a companion robot, a shopping guide robot, etc.; a robot used in the medical field, such as a surgical robot, a nursing robot, etc.

[0040] In addition, the local network is a local area network. Data transmission within the local network does not need to pass through multiple forwarding nodes and has the characteristics of high transmission efficiency.

[0041] In addition, the robot may be provided with at least one of the following sensors: an RGB camera for collecting image data or video, a depth camera for obtaining depth information (such as structured light) of an object, a laser radar for obtaining three-dimensional point cloud data through laser ranging, a stereoscopic vision system for using multiple cameras to obtain depth information, an ultrasonic sensor for measuring distance by emitting sound waves and receiving echoes, an infrared sensor for measuring distance or detecting objects by emitting infrared rays and receiving reflected infrared light signals, a capacitive sensor for detecting approaching objects by measuring changes in the electric field between the object, a force sensor for measuring the force acting on the robot, a torque sensor for measuring torque on a robot joint or other rotating parts, an accelerometer for measuring the acceleration of the robot in all directions, a gyroscope for measuring the angular velocity of rotation of the robot, a temperature and humidity sensor for measuring the temperature / humidity of the environment or the robot, a global positioning system for measuring the position of the robot, and an optical sensor for sensing the light intensity of the environment. The number of any sensor on the robot may be one or more, and the embodiments of the present application are not limited thereto.

[0042] On the basis that the robot is provided with one or more of the above sensors, the robot can collect data through the sensors and upload the data to the private cloud device for further processing. Optionally, the data includes but is not limited to: RGBD data, fisheye camera data, laser radar data, robot joint angle data.

[0043] Furthermore, the private cloud device can call the corresponding service instance to perform frame alignment on the data according to the sensor distribution information corresponding to the robot to obtain the data processing result. The sensor distribution information corresponding to the robot can indicate the categories of the sensors set on the robot and the corresponding setting positions.

[0044] Specifically, the robot may include but is not limited to: a bipedal robot for simulating human bipedal walking and equipped with multiple joints and sensors, a wheeled robot for moving on wheels and equipped with drive wheels, steering wheels and sensors, a tracked robot for moving using tracks, a multi-legged robot for simulating animal movements, a flying robot, etc.

[0045] Different robots may correspond to different sensor distribution information. For example, the sensors of bipedal robots are generally installed on the head, chest, legs, etc., while the sensors of wheeled robots are usually installed around the wheeled robots. In other words, the types of data acquired by different robots may be different, that is, the data that needs to be aligned is different. Therefore, the private cloud device can call the corresponding service instance according to the sensor distribution information. The corresponding service instance can be dedicated to frame alignment of the robot's data to obtain data processing results. The service instance can include corresponding data alignment algorithms, sensor synchronization algorithms and other algorithms.

[0046] Among them, because there may be time delays in the data collection of different sensors, it is necessary to align the data from different sensors to the same timestamp through the above service instance and then store them. Based on this, multiple data with the same timestamp can be processed. The processing here can refer to operations such as generating environmental images, constructing three-dimensional space, robot position judgment, and robot state judgment. The processed data can be applied to subsequent decisions. For example, based on the result of the robot position judgment, a robot scheduling task for adjusting the robot's posture is sent to the robot. Among them, the robot scheduling task can control the robot to perform specific operations, such as grasping, avoiding obstacles, adjusting posture, etc.

[0047] The data processing method provided in the embodiment of the present application can receive data transmitted by the robot through the local network, and call the corresponding service instance to frame align the data according to the sensor distribution information corresponding to the robot, so as to obtain the data processing result and reduce the computing pressure of the robot.

[0048] In some embodiments, the private cloud device can label the data processing results according to labeling rules to obtain labeled data; wherein the labeling rules are used to indicate at least one data labeling dimension; the labeled data is uploaded to the public cloud device so that the public cloud device performs a model training task based on the labeled data; wherein the public cloud device and the private cloud device transmit data through an external network.

[0049] In this embodiment, in order to further improve data utilization and save computing power of private cloud devices, private cloud devices can use classification models, recognition models and other models to label data processing results according to labeling rules, and upload the labeled data to public cloud devices through an external network, so that the public cloud devices can perform model training tasks with low urgency based on the labeled data.

[0050] Among them, optionally, the data annotation dimensions indicated by the annotation rules can be: the position of the detected object, the category of the detected object, the attribute of the detected object, the application scenario, the semantic label, the obstacle position, the drivable area, the road markings, the traffic signs, etc. The specific data annotation dimensions to be used depend on actual needs.

[0051] In addition, optionally, the private cloud device can also directly upload the data processing results to the public cloud device, so that the public cloud device can perform data annotation on the data processing results.

[0052] In some embodiments, the private cloud device annotates the data processing results according to the annotation rules. Before obtaining the annotated data, the private cloud device can also display the data processing results on the user review interface; when an adjustment operation on the data processing results is detected, the data processing results are updated.

[0053] In this embodiment, in order to ensure the accuracy of the data processing results, the private cloud device can display the data processing results on the user review interface. When an adjustment operation on the data processing results is detected, the data processing results can be updated. The updated data processing results have higher accuracy than the data processing results before the update. Annotating the updated data processing results can obtain more accurate annotation results. Among them, the user review interface is a visual interface, and the data processing results can be displayed directly in the visual interface or in a specific format in the visual interface. This embodiment of the application does not limit this.

[0054] In some embodiments, the data is obtained by the robot compressing the original data collection results based on a compression rule, and the compression rule is used to define the compression method corresponding to different data types.

[0055] In this embodiment, in order to improve data transmission efficiency, the robot can compress the original data collection results based on compression rules to obtain data that can be efficiently transmitted to the private cloud device. After receiving the data, the private cloud device can parse the data based on the agreed compression rules to obtain the original data before the data is compressed, which is conducive to the private cloud device to perform more accurate and effective data processing operations.

[0056] Specifically, the raw data acquisition result refers to the data collected by various sensors set on the robot. As an optional embodiment, the raw data acquisition result includes at least one of the following: RGBD data, fisheye camera data, laser radar data, and robot joint angle data. Among them, RGBD data refers to data containing color images (i.e., RGB images) and depth information (D). The depth information is used to indicate the distance between each pixel and the camera. The color image can be collected by a conventional camera, and the depth information can be collected by a depth camera. RGBD data can be used to identify obstacles in the environment and grasp objects. Among them, the fisheye camera data is data collected by an ultra-wide-angle fisheye lens. Compared with conventional cameras, the fisheye camera data contains a wider field of view and can be applied to robot vision and other aspects. Laser radar data is the distance data obtained by laser radar using laser to scan objects and measure the time of reflected laser. In terms of form, laser radar data is a three-dimensional point cloud data. Each point in the three-dimensional point cloud data comprehensively represents the spatial position of the surface of the object. Laser radar data can be used to construct three-dimensional modeling within the scene. The robot joint angle data is data collected by encoders used to measure the rotation angles of mechanical parts. The robot joint angle data represents the angle information of one or more joints of the robot. Controlling the robot joint angle data can help the robot better complete specific tasks, such as obstacle avoidance and grasping tasks.

[0057] Specifically, the compression rule is used to define at least one data compression method for various types of raw data collection results. As an optional embodiment, the compression rule defines at least one of the following compression methods: RGBD data corresponding to a 10-fold compression rate is obtained by H.265 encoding compression, fisheye camera data corresponding to a 10-fold compression rate is obtained by H.265 encoding compression, laser radar data corresponding to a 75% compression rate is obtained by compression based on a lossless compression algorithm, and robot joint angle data corresponding to a 5-fold compression rate is obtained by compression according to the HDF5 format. Among them, H.265 encoding is a video compression standard. Compressing RGBD data and fisheye camera data based on H.265 encoding can reduce the amount of data transmission and improve data transmission efficiency. The lossless compression algorithm is an algorithm that can restore the data to its original state after compression. Since the laser radar data is point cloud data containing spatial coordinates (X, Y, Z), and the loss of any data in the spatial coordinates will result in the inability to reconstruct the three-dimensional environment, the use of a lossless compression algorithm for the laser radar data can ensure the transmission accuracy of the laser radar data. The HDF5 format is a data storage format. Since robot joint angle data is usually stored as an array, the HDF5 format can be used to compress robot joint angle data.

[0058] In some embodiments, the data is obtained by the robot verifying the data compression result based on verification rules, and the verification rules are used to limit at least one of the following verification dimensions: data integrity, data quality, system abnormality, interaction environment abnormality, and task execution abnormality.

[0059] In this embodiment, in order to improve data quality, ensure data integrity, and mine the potential value in the data, the robot can verify the original data collection results based on the verification rules. The data that passes the verification can be stored as data to be uploaded, and the data that fails the verification can be discarded or fed back to the manual processing flow, thereby improving the quality of data transmitted to the private cloud device and avoiding the impact of data that fails the verification on the data alignment efficiency of the private cloud device. In addition, the verification can also be used to mine potential valuable data in the original data collection results, such as environmental interference factor data.

[0060] Optionally, the verification rules are used to limit the verification dimensions including: data integrity, data quality, system abnormality, interaction environment abnormality, and task execution abnormality.

[0061] The data integrity check as a verification dimension includes the following items: whether the directory structure of the data file is complete, whether it contains the specified key files, whether the file size and format meet the specifications, whether the file is not damaged, whether the message count is correct, whether the timestamp is continuous and valid, and whether the data format meets the specifications.

[0062] As a verification dimension, the data quality indicator checks the following items: whether the H.265 video file is complete, whether the frame rate loss rate meets expectations, whether the frame interval time is standardized, whether the accuracy of the PTP timestamp meets the specifications, whether the synchronization alignment between multiple cameras is normal, whether the robot arm data is complete, whether the joint angle is within the specified range, whether the robot arm's motion trajectory is continuous, whether the relevant data of the gripper is valid, and whether the force feedback data is valid.

[0063] The system abnormality check as a verification dimension includes the following items: whether the operation status of each node is normal, whether the execution status of the algorithm is normal, and whether the resource usage of the system is normal.

[0064] The following items are checked for abnormal interaction environment as a verification dimension: whether there is occlusion in the point cloud data, whether there are occlusions in the camera field of view, and whether there are environmental interference factors (such as lighting changes, object reflections, etc.).

[0065] The following items are checked for abnormality in task execution as verification dimensions: whether the quality of the motion trajectory meets expectations, the success rate of the grasping task, and the degree of task completion.

[0066] The data obtained by verifying the raw data collection results based on the verification rules can be stored in a unified data format for unified management and access. For example, the execution action of the robot can be recorded in the root directory / action, and the state of the robot can be recorded in the root directory / state. Furthermore, the action of the robot actuator is recorded through / action / effector, the force of the actuator is recorded through / action / effector / force, the position of the actuator is recorded through / action / effector / position, the timestamp of the actuator action is recorded through / action / effector / timestamp, the action of the robot head is recorded through / action / head, the effort of the head is recorded through / action / head / effort, the position of the head is recorded through / action / head / position, the speed of the head is recorded through / action / head / velocity, the action of the robot system is recorded through / action / robot, the posture change of the robot is recorded through / action / robot / angular, and the timestamp of the actuator action is recorded through / action / ion / robot / orientation records the robot's orientation, / action / robot / velocity records the robot's movement speed, / state / effector records the actuator's state, / state / effector / force records the actuator's current force, / state / effector / position records the actuator's current position, / state / effector / timestamp records the state timestamp, / state / robot records the robot's state, / state / robot records the robot's current direction through / state / robot / orientation, / state / robot / position records the robot's current position, and / state / robot / timestamp records the robot's state timestamp.

[0067] In some embodiments, the private cloud device includes a data processing module and an object storage module; the private cloud device receives data transmitted by the robot through the local network, and the data is obtained by collecting sensors, including: when a token request sent by the robot is received, a token for identifying a data collection cycle is sent to the robot through the data processing module, so that the robot collects raw data based on the token, compresses and verifies the raw data into the data, and then uploads it to the object storage module, and returns the upload status of the data to the data processing module; when the upload status is received, the data processing module is triggered to pull the data from the object storage module.

[0068] In this embodiment, in order to improve the accuracy of data processing and the security of data transmission, the private cloud device can send a token (episode token) for identifying the data collection cycle to the robot through a data processing module (i.e., data platform) for collecting, processing, storing and distributing data when receiving a token request (get episode token) sent by the robot or the human-computer interaction interface of the robot. After receiving the token, the robot can start to perform a cycle of data collection and store the original data. Then, the robot can perform frame rate verification on the original data to ensure that the speed and frequency of data collection meet expectations and avoid frame loss or discontinuous data collection, where the frame rate refers to the acquisition speed or refresh rate of data or video frames. Furthermore, the robot can also calculate the hash value of the original data based on a hash algorithm (such as MD5) and compress and verify the original data into the data and then upload it to the object storage module (i.e., OSS), and the data processing module returns the upload status for the data (such as upload success, upload failure, etc.) and the above hash value, where the object storage module is used to store unstructured data. In the case of receiving the upload status, the data processing module feedbacks to the robot the status indicating whether the upload status is successfully uploaded, and the object storage module pulls the data. The data processing module can also verify the correctness of the data based on the aforementioned hash value, and after successful verification, it can frame align the data and mark the data processing results. In addition, optionally, the object storage module can also review the marking accuracy of the data processing results based on the marking rules.

[0069] In some embodiments, the private cloud device calls the corresponding service instance to perform frame alignment on the data according to the sensor distribution information corresponding to the robot to obtain the step of data processing results, including: calling the corresponding service instance according to the sensor distribution information corresponding to the robot and allocating corresponding computing resources to the service instance, so that the service instance determines the data list corresponding to each data acquisition unit in the robot based on the data, and checks the timestamp consistency of each group of data frames in turn through the created current frame group to obtain the data processing results; wherein, the data frames of any group of data frames are taken from each data list respectively.

[0070] In this embodiment, in order to achieve accurate data alignment, the private cloud device can call the corresponding service instance and allocate corresponding computing resources (such as CPU resources, memory resources, etc.) to the service instance according to the sensor distribution information corresponding to the robot, so that the service instance determines the data list corresponding to each data acquisition unit (such as a camera) in the robot based on the data, and checks the timestamp consistency of each group of data frames in turn through the created current frame group to obtain the data processing result.

[0071] The service instance determines the data list corresponding to each data acquisition unit in the robot based on the data, and checks the timestamp consistency of each group of data frames in turn through the created current frame group to obtain the data processing result. The specific method can be: initialize the data acquisition unit index, such as camera_index = 0. Initialize the data list of each data acquisition unit, such as camera_files = ['Camera1_data', 'Camera2_data', 'Camera3_data', 'Camera4_data']. Set the frame index and frame information dictionary, such as frame_index = 0, frame_info = {}.

[0072] Furthermore, the data in the data list of each data acquisition unit can be processed in groups by loop execution, specifically: create a current frame group, such as current_frames = [], and traverse the data list of each data acquisition unit to add the data currently to be processed to the current_frames list, and then call the single_frame (current_frames) function to check whether each frame in the current_frames list is aligned in time order and whether each data acquisition unit has frame loss or delay. If each data is aligned in time order and each data acquisition unit has no frame loss or delay, then initialize the data list and data processing results of each data acquisition unit processed this time, such as current_files = ['Camera1_data_1', 'Camera2_data_1', 'Camera3_data_2', 'Camera4_data_1'], current_frame_info = {'cameras_with_data': ['Camera1', 'Camera2', 'Camera4'], 'cameras_without_data': ['Camera3']}. Then, based on the above data processing results, the data list and data acquisition unit index of each data acquisition unit are updated, such as camera_index=1, current_files, so as to continue processing the next set of data, and the frame information dictionary is updated, such as frame_info[frame_index]={'cameras_with_data':['Camera1','Camera2','Camera4'],'cameras_without_data':['Camera3']}. Then, the data processing results are added to the frame information dictionary and the group status is recorded, such as frame_info={0:{'cameras_with_data':['Camera1','Camera2','Camera4'],'cameras_without_data':['Camera3']},...}. Then, the data list information is added to the frame information dictionary file list, such as frame_index_files = [['Camera1_data_1', 'Camera2_data_1', 'Camera3_data_2', 'Camera4_data_1']]. By looping the above steps, the data in the data list of each data acquisition unit can be processed in groups in turn until all the data in the cycle is processed.

[0073] See also Figure 3One embodiment of the present application also provides a data processing device. The data processing device may include: a data receiving unit, used to receive data transmitted by the robot through a local network, the data being acquired through sensors; a data processing unit, used to call a corresponding service instance to perform frame alignment on the data according to the sensor distribution information corresponding to the robot, to obtain a data processing result.

[0074] In this embodiment, the specific functions and effects achieved by the data processing device can be explained with reference to other embodiments of the present application and will not be repeated here.

[0075] The embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the processor implements the data processing method as described above.

[0076] The embodiment of the present application also provides a computer program product including instructions, and when the computer program product is executed by a processor, the data processing method as described above is implemented.

[0077] See also Figure 4 The present description embodiment may provide a computer device, the computer device comprising: a memory, and one or more processors communicatively connected to the memory; the memory stores instructions executable by the one or more processors, the instructions are executed by the one or more processors, so that the one or more processors implement the data processing method as described above.

[0078] In some embodiments, the computer device may include a processor, a non-volatile storage medium, an internal memory, a communication interface, a display device, and an input device connected by a system bus. The non-volatile storage medium may store an operating system and related computer programs.

[0079] It should be understood that the specific examples in this article are only intended to help those skilled in the art to better understand the embodiments of the present application, rather than to limit the scope of the present invention.

[0080] It can be understood that in the various implementations of the present application, the size of the serial number of each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the implementation methods of the present application.

[0081] It can be understood that the various embodiments described in this application can be implemented individually or in combination, and the embodiments of this application are not limited to this.

[0082] Unless otherwise stated, all technical and scientific terms used in the embodiments of the present application have the same meaning as those generally understood by those skilled in the art of the technical field of the present application. The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the scope of the present application. The term "and / or" used in the present application includes any and all combinations of one or more related listed items. The singular forms of "a kind of", "above" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.

[0083] It can be understood that the processor of the embodiment of the present application can be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method implementation can be completed by the hardware integrated logic circuit or software instructions in the processor. The above processor can be a general processor, a digital signal processor (DigitalSignal 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 methods, steps and logic block diagrams disclosed in the embodiment of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to perform, or the hardware and software modules in the decoding processor are combined and executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0084] It is understood that the memory in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memory. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (programmable ROM, PROM), an erasable programmable read-only memory (erasable PROM, EPROM), an electrically erasable programmable read-only memory (EEPROM) or a flash memory. The volatile memory may be a random access memory (RAM). It should be noted that the memory of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0085] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0086] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method implementation methods and will not be repeated here.

[0087] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device implementation described above is only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple 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, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0088] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0089] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0090] If the function 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 can be essentially or partly embodied in the form of a software product that contributes to the prior art, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0091] The above is only a specific implementation of the present application, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A data processing method, characterized in that: The method is applied to a private cloud device, the private cloud device and the robot perform data transmission via a local network, and the method comprises: Receiving data transmitted by the robot through the local network, wherein the data is collected by a sensor; According to the sensor distribution information corresponding to the robot, the corresponding service instance is called to perform frame alignment on the data to obtain a data processing result.

2. The method according to claim 1, characterized in that The data is obtained by the robot compressing original data based on a compression rule, and the compression rule is used to define the compression method corresponding to different data types.

3. The method according to claim 2, characterized in that The data is obtained by the robot verifying the data compression result based on the verification rules, and the verification rules are used to limit at least one of the following verification dimensions: data integrity, data quality, system abnormality, interaction environment abnormality, and task execution abnormality.

4. The method according to claim 1, characterized in that The step of calling a corresponding service instance to perform frame alignment on the data according to the sensor distribution information corresponding to the robot to obtain a data processing result includes: According to the sensor distribution information corresponding to the robot, the corresponding service instance is called and the corresponding computing resources are allocated to the service instance, so that the service instance determines the data list corresponding to each data acquisition unit in the robot based on the data, and checks the timestamp consistency of each group of data frames in turn through the created current frame group to obtain the data processing result; wherein, the data frames of any group of data frames are taken from each data list respectively.

5. The method according to claim 1, characterized in that The private cloud device includes a data processing module and an object storage module; The step of receiving data transmitted by the robot through the local network, wherein the data is collected by a sensor, comprises: In the case of receiving a token request sent by the robot, the data processing module sends a token for identifying a data collection period to the robot, so that the robot collects raw data based on the token, compresses and verifies the raw data into the data, and then uploads it to the object storage module, and returns the upload status of the data to the data processing module; When the upload status is received, the data processing module is triggered to pull the data from the object storage module.

6. The method according to claim 1, characterized in that The method further comprises: Annotating the data processing result according to the annotation rule to obtain annotated data; wherein the annotation rule is used to indicate at least one data annotation dimension; The annotated data is uploaded to a public cloud device, wherein the public cloud device and the private cloud device perform data transmission via an external network.

7. The method according to claim 6, characterized in that Before the step of labeling the data processing result according to the labeling rules to obtain the labeled data, the method further includes: Display the data processing results on the user review interface; When an adjustment operation on the data processing result is detected, the data processing result is updated.

8. A data processing system, characterized in that: The data processing system includes a robot and a private cloud device, and the robot and the private cloud device perform data transmission via a local network, wherein: The robot is used to collect data through sensors and transmit the data to the private cloud device through the local network; The private cloud device is used to call the corresponding service instance to perform frame alignment on the data according to the sensor distribution information corresponding to the robot to obtain the data processing result.

9. The system according to claim 8, characterized in that The data processing system also includes a public cloud device, wherein: The public cloud device is used to receive the data processing result sent by the private cloud device.

10. A computer device, characterized in that: The computer device includes a memory and a processor, wherein the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor to implement the data processing method according to any one of claims 1 to 7.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one computer program, and when the at least one computer program is executed by a processor, it can implement the data processing method according to any one of claims 1 to 7.

12. A computer program product, characterized in that The computer program product is used to implement the data processing method according to any one of claims 1 to 7.

Images