Simulation data processing method, device and system of sensor

Abstract

The invention relates to a simulation data processing method, device and system of a sensor, and the method comprises the steps: controlling a real sensor of a robot to carry out the data collection of a target object through different target parameters, and obtaining the real sensor data corresponding to different target parameters; screening effective data related to the target object in the real sensor data to obtain effective real sensor data corresponding to different target parameters; based on effective real sensor data corresponding to different target parameters, fitting to obtain a mapping relation between the target parameters of the target object and the sensor data; the mapping relation is used for taking simulation parameters of the simulation sensor of the robot in the simulation environment as input and outputting simulation sensor data including noise of the target object. Since the mapping relation can reflect the fluctuation rule of the sensor in a real environment, the simulation sensor data including noise corresponding to the simulation parameters is predicted based on the mapping relation, so that the simulation sensor data has authenticity, reliability and accuracy.

Description

Technical Field

[0001] The present application relates to the field of robotics technology, and in particular to a sensor simulation processing method, device, system and computer equipment. Background Art

[0002] With the development of robot technology, robots have greatly facilitated people's lives. In the design stage of robot products, in order to produce qualified robot products faster and better, it is usually necessary to simulate various working conditions of the robot and test the product performance.

[0003] Usually, the sensor data used in simulation is based on the return value of the simulated sensor plus Gaussian noise. However, in the real world, affected by various factors such as material and lighting, sensor data often has various forms of noise. With traditional technology, the noise does not take into account the factors of the real environment. There is a big difference between the sensor data in the simulation and the data in the real treadmill, resulting in a big difference between the performance of the robot in the simulation and the real robot. Some problems that occur in the real machine test during the robot development process cannot be reproduced well in the simulation environment, which affects the efficiency and performance of the robot development. Summary of the invention

[0004] Based on this, it is necessary to provide a simulation sensor data processing method, device, system, computer equipment, computer-readable storage medium and computer program product that can improve the reliability of simulation data and thereby improve the efficiency and performance of robot development in response to the above-mentioned technical problems.

[0005] In a first aspect, the present application provides a method for processing simulation data of a sensor, the method comprising:

[0006] The real sensor of the robot is controlled to collect data of the target object with different target parameters, and the real sensor data corresponding to the different target parameters are obtained;

[0007] Filtering valid data related to the target object in the real sensor data to obtain valid real sensor data corresponding to the different target parameters;

[0008] Based on the effective real sensor data corresponding to the different target parameters, a mapping relationship between the target parameters and the sensor data of the target object is fitted; the mapping relationship is used to take the simulation parameters of the simulated sensor of the robot in the simulation environment as input, and output the simulated sensor data including noise for the target object.

[0009] In a second aspect, the present application provides a sensor simulation data processing device, comprising:

[0010] An acquisition module is used to control the real sensor of the robot to collect data of the target object with different target parameters to obtain the real sensor data corresponding to the different target parameters;

[0011] A screening module, used for screening the valid data related to the target object in the real sensor data, and obtaining the valid real sensor data corresponding to the different target parameters;

[0012] A prediction module is used to fit the mapping relationship between the target parameters and sensor data of the target object based on the effective real sensor data corresponding to the different target parameters; the mapping relationship is used to take the simulation parameters of the simulated sensor of the robot in the simulation environment as input, and output the simulated sensor data including noise for the target object.

[0013] In a third aspect, the present application provides a sensor simulation data processing system, comprising:

[0014] A data acquisition device, used to control the real sensor of the robot to collect data on the target object with different target parameters, and obtain the real sensor data corresponding to the different target parameters;

[0015] A host computer is connected to the data acquisition device and is used to screen the valid data related to the target object in the real sensor data to obtain the valid real sensor data corresponding to the different target parameters; based on the valid real sensor data corresponding to the different target parameters, a mapping relationship between the target parameters and the sensor data of the target object is obtained by fitting; the mapping relationship is used to take the simulation parameters of the simulated sensor of the robot in the simulation environment as input, and output the simulated sensor data including noise for the target object.

[0016] In one embodiment, the data acquisition device comprises:

[0017] Sensor motion platform, used to load real sensors;

[0018] The object loading platform is used to load the target object; the motion platform is arranged opposite to the object loading platform;

[0019] An environmental adjustment device, used to adjust environmental parameters;

[0020] a controller, electrically connected to the sensor motion platform, the object loading platform and the environment adjustment device, for controlling at least one of the sensor motion platform, the object loading platform and the environment adjustment device to adjust according to the target parameter;

[0021] The host computer is directly or indirectly connected to the real sensor to obtain real sensor data.

[0022] In one embodiment, the target parameters include: sensor motion parameters, the relative distance between the sensor and the target object, and environmental parameters; the sensor motion parameters include the motion trajectory and motion speed of the sensor motion platform; the controller is used to control the environmental adjustment device to provide a collection site environment based on the environmental parameters; control the object loading platform and the sensor motion platform to maintain the relative distance, and the target object is set on the object loading platform; control the sensor motion platform to move with the motion trajectory and the motion speed, and control the real sensor to collect data on the target object during the movement to obtain corresponding real sensor data.

[0023] In a fourth aspect, the present application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the methods of the above embodiments when executing the computer program.

[0024] In a fifth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method of each of the above embodiments are implemented.

[0025] In a sixth aspect, the present application provides a computer program product, including a computer program, which implements the steps of the methods of the above embodiments when executed by a processor.

[0026] The simulation data processing method, device and system of the above-mentioned sensor obtains the mapping relationship between the target parameter and the sensor data by discretely collecting the real sensor data under different target parameters, mining the fluctuation law of the sensor in the real environment, and applying the mapping relationship to the simulation environment. The simulation parameters of the robot's simulation sensor in the simulation environment are used as input to output the simulated sensor data including noise for the target object. Since the mapping relationship can reflect the fluctuation law of the sensor in the real environment, the simulated sensor data including noise corresponding to the simulation parameters is predicted based on the mapping relationship, so that the simulated sensor data has authenticity, reliability and accuracy, thereby improving the simulation effect and the efficiency of robot product development. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a structural block diagram of a simulation data processing system for a sensor in one embodiment;

[0028] Figure 2 A schematic diagram of a flow chart of a method for processing simulation data of a sensor in one embodiment;

[0029] Figure 3A flowchart of the steps of acquiring real sensor data corresponding to different target parameters for a target object using a real sensor controlling a robot in one embodiment;

[0030] Figure 4 A schematic flow chart of steps for controlling a real sensor to collect data of a target object with target parameters to obtain corresponding real sensor data according to an embodiment;

[0031] Figure 5 It is a structural block diagram of a simulation data processing device for a sensor;

[0032] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying 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.

[0034] The present application provides a sensor simulation processing method, which can be applied to Figure 1 The simulation environment shown in the figure may include: a simulation platform 10 and a simulation data processing system 20. The simulation data processing system 20 includes a data acquisition device 21 and a host computer 22, which are electrically connected.

[0035] The data acquisition device 21 is used to control the real sensor of the robot to collect data on the target object with different target parameters to obtain the real sensor data corresponding to the different target parameters.

[0036] The host computer 22 is connected to the data acquisition device and is used to screen the real sensor data and the valid data related to the target object to obtain the valid real sensor data corresponding to the different target parameters; based on the valid real sensor data corresponding to the different target parameters, a mapping relationship between the target parameters and the sensor data of the real sensor to the target object is obtained by fitting; the mapping relationship is used to predict the simulated sensor data of the target object based on the parameters of the sensor in the simulation environment.

[0037] The above-mentioned sensor simulation data processing system obtains the mapping relationship between the target parameters and the sensor data by discretely collecting real sensor data under different target parameters, mining the fluctuation law of the sensor in the real environment, and applying the mapping relationship to the simulation environment. It takes the simulation parameters of the robot's simulation sensor in the simulation environment as input and outputs the sensor data including noise for the target object. Since the mapping relationship can reflect the fluctuation law of the sensor in the real environment, the simulation sensor data including noise corresponding to the simulation parameters is predicted based on the mapping relationship, so that the simulation sensor data has authenticity, reliability and accuracy, which can improve the simulation effect and the efficiency of robot product development.

[0038] In another embodiment, the data acquisition device 21 includes: a sensor motion platform 211 for loading a real sensor, an object loading platform 212 for loading a target object, an environment adjustment device 213 and a controller 214 for adjusting environmental parameters. The sensor motion platform 211 and the object loading platform 212 are arranged relative to each other, and the distance between the two can be adjusted according to the target parameters. The motion platform 211 and the object loading platform 212 are both electrically connected to the controller 214, and the movement of the two is controlled by the controller. The real sensor is electrically connected to the host computer 22, and the collected real sensor data is sent to the host computer 22, or the real sensor is electrically connected to the controller 214, and the collected real sensor data is sent to the host computer 22 through the controller 214. The controller 214 is electrically connected to the host computer 22, and the controller sends the relevant parameters of the sensor motion platform 211 and the object loading platform 212 to the host computer. During the test, the sensor motion platform 211 is first moved to the specified position, and then the motion trajectory of the sensor motion platform is planned. The planned platform motion trajectory is designed based on the sensor data acquisition system, taking into account the efficiency of automated testing and the range of data point acquisition. Through simulation data processing 20, a universal sensor data acquisition platform can be provided, which can be adapted to various types of sensors and complete sensor data acquisition of various different types of sensors. The simulation data processing system can collect data from different sensors at different positions, speeds, illumination, and types of objects being measured.

[0039] Based on the above application environment, the present application provides a sensor simulation data processing method, which is applied to Figure 1 The host computer 22 shown in FIG. Figure 2 As shown, the following steps are included:

[0040] Step 202 , controlling the real sensor of the robot to collect data on the target object with different target parameters, and obtaining real sensor data corresponding to the different target parameters.

[0041] Among them, the real sensor is the sensor that needs to be installed on the robot product. The sensor plays a key role in the robot. By effectively simulating the sensor, the performance of the sensor can be detected and the robot test can be improved. In this embodiment, the type of sensor is not limited, and can include at least one of radar, wall sensor, image sensor and IMU sensor. That is, the simulation data processing method of the sensor of the present application can be used to realize the simulation data processing of the above-mentioned sensor.

[0042] The target object is an object that appears in the robot application environment. The target object setting can be different for different robots. For example, for a sweeping robot, the corresponding target object is an object that often appears in an indoor environment, such as a sofa, tile, mirror, and cabinet. Considering that the noise of the sensor is usually easily affected by factors such as material, color, glossiness, and lighting, and has little correlation with factors such as shape, the target object can also be the surface material of an object that appears frequently in the robot application environment. For example, the surface material of a sofa, the surface material of a cabinet, etc.

[0043] The target parameters are used to simulate different working conditions of the robot in the real environment. The target parameters can include the multi-dimensional parameters of the robot in the real sensor, including sensor motion parameters, sensor acquisition parameters, environmental parameters, and the relative distance between the sensor and the target object.

[0044] Among them, the sensor motion parameters are related to the robot motion in the actual scene, which reflects the operation of the robot. The sensor motion parameters can reflect the operation of the sweeping robot on the ground, for example, translation and rotation at different distances or speeds.

[0045] Sensor acquisition parameters refer to the parameters used by the sensor when collecting data, such as the focal length of the image sensor, the laser radar line distance, the incident angle, etc.

[0046] Environmental parameters refer to the parameters of the robot's environment, which may include light, temperature, humidity, etc.

[0047] The sensor has a wide acquisition range and needs to collect data in segments. For example, the target object can be fixed first, and then the distance between the sensor and the target object can be adjusted to obtain sensor data at different distances and evaluate the impact of distance on the real sensor data of noise.

[0048] In order to collect as much real sensor data as possible, in this embodiment, the target parameters can be changed to obtain different target parameters, and the real robot can be controlled to collect data on the target object with different target parameters to obtain real sensor data corresponding to different target parameters.

[0049] Step 204 , screening the valid data related to the target object in the real sensor data to obtain the valid real sensor data corresponding to different target parameters.

[0050] Among them, valid data refers to data in real sensor data that meets the requirements and can be used to mine the noise law of the sensor under different target parameters. In one embodiment, some data of non-target objects may be collected in the real sensor data, and these data cannot reflect the noise law of the sensor under different target parameters. Therefore, the data irrelevant to the target object in the real sensor data can be screened and then deleted to obtain valid real sensor data.

[0051] Specifically, the data points of the target object in the real sensor coordinate system in the real sensor data can be determined according to the relative distance between the real sensor and the target object and the coordinates of the target object, and the valid real sensor data in the real sensor data can be obtained. In addition, a box can be installed during data collection to cover the measurement range where invalid data is located to reduce the workload of data processing.

[0052] Step 206, based on the effective real sensor data corresponding to different target parameters, fit the mapping relationship between the target parameters and the sensor data of the target object; the mapping relationship is used to take the simulation parameters of the robot's simulation sensor in the simulation environment as input, and output the simulated sensor data including noise for the target object.

[0053] Specifically, based on the analysis of valid real data corresponding to different target parameters, the mapping relationship between the target parameters and the sensor data is mined. In one embodiment, the mapping relationship between the target parameters and the sensor data can be a mapping relationship between the sensor's target parameters for the target object and the sensor data including noise. In one embodiment, the mapping relationship between the target parameters and the sensor data can also be a mapping relationship between the sensor's target parameters for the target object and noise data.

[0054] Among them, data mining can adopt the method of machine learning. For example, unsupervised learning can be performed based on different target parameters and corresponding valid real data to mine the mapping relationship between target parameters and sensor data. Among them, the random forest algorithm can be used for unsupervised learning.

[0055] In one embodiment, the mapping relationship between the target parameters of the target object obtained by fitting and the sensor data is sent to the host computer. After the host computer obtains the mapping relationship between the target parameters of the target object and the sensor data of the real sensor, the mapping relationship is configured onto the simulation platform. The simulation platform is used to simulate the robot running on the treadmill in the real environment. According to the simulation parameters of the simulation environment, based on the mapping relationship between the target parameters of the target object and the sensor data of the sensor, the simulation sensor data including noise of the target object is predicted. Since the simulation sensor data is obtained based on the mapping relationship between the target parameters of the target object and the sensor data of the real sensor, which conforms to the fluctuation law of the noise of the real sensor under different target parameters, the authenticity of the simulation sensor data is improved, and thus the treadmill running effect in the simulation is improved.

[0056] The method of the present application mines the mapping relationship between the target parameters and the sensor data of the sensor by collecting discrete real sensor data under different target parameters, and applies the mapping relationship to the simulation environment. Taking the parameters of the robot in the simulation environment as the input, the simulation sensor data including noise of the target object is output. Since the mapping relationship can reflect the fluctuation law of the sensor in the real environment, and the simulation sensor data including noise corresponding to the simulation parameters is predicted based on the mapping relationship, the sensor data in the simulation has authenticity, reliability and accuracy, and thus the treadmill running effect in the simulation can be improved, and the development efficiency of the robot product can be enhanced.

[0057] In another embodiment, the real sensor of the robot is controlled to collect data from the target object with different target parameters, and the real sensor data corresponding to different target parameters is obtained, as Figure 3 shown, including:

[0058] Step 302, obtain the target parameters of the real sensor of the robot.

[0059] Specifically, at initialization, the initial target parameters can be obtained. The target parameters can include multi-dimensional parameters of the robot in the real sensor, including sensor motion parameters, sensor acquisition parameters, environmental parameters, and the relative distance between the sensor and the target object, etc.

[0060] Step 304, control the real sensor to collect data from the target object with the target parameters to obtain the corresponding real sensor data.

[0061] Specifically, control the real sensor to run and collect data with at least one of the sensor motion parameters, sensor acquisition parameters, environmental parameters, and the relative distance between the sensor and the target object to obtain the real sensor data.

[0062] Step 306, changing the target parameters, and returning to execute the step of obtaining the target parameters of the real sensor of the robot, to obtain the real sensor data corresponding to different target parameters.

[0063] The target parameter has multiple dimensions, for example, it may include at least one of a sensor motion parameter, a sensor acquisition parameter, an environmental parameter, and a relative distance between the sensor and the target object. In this embodiment, by changing the target parameter and controlling the sensor to collect target data under the changed target parameter, real sensor data corresponding to different target parameters can be obtained.

[0064] The change method may be to change the target parameters of one dimension at a time, fix the target parameters of other dimensions, and then collect sensor data. After traversing the optional parameters of each dimension, continue to change the target parameters of the next dimension. After completing the traversal of the target parameters of all dimensions, the real sensor data corresponding to different target parameters can be obtained.

[0065] Taking the distance between the sensor and the target object as an example, the test distance range is 0-9 meters. The test starts from 0 meters, and each change can increase the distance by 0.5 meters. For each distance, the sensor motion parameters, sensor acquisition parameters and environmental parameters can be changed in turn. After completing the traversal of all parameters, continue to change the distance and repeat the above process to obtain the real sensor data corresponding to different target parameters.

[0066] In this embodiment, by setting the target parameter-collection-changing the target parameter-collection cycle, real sensor data corresponding to different parameters can be obtained, enriching the test data used for mapping relationship mining.

[0067] In another embodiment, the target parameter includes at least one of a sensor motion parameter, a distance to a target object, a sensor acquisition parameter, and an environmental parameter.

[0068] Among them, the sensor motion parameters are related to the robot motion in the actual scene, and they reflect the robot's operating status. By taking the sensor operation data as one of the dimensions of the target parameters, it is possible to consider a variety of sensor motion parameters and explore the impact of different sensor operation parameters on the sensor data fluctuation pattern.

[0069] Sensor acquisition parameters refer to the parameters used by sensors when collecting data. By taking sensor acquisition parameters as one of the dimensions of target parameters, it is possible to consider multiple sensor parameters and explore the impact of different sensor parameters on the fluctuation pattern of sensor data.

[0070] Environmental parameters refer to the parameters of the robot's environment, which can include light, temperature, humidity, etc. By taking environmental parameters as one of the dimensions of the target parameters, it is possible to consider a variety of environmental variables and explore the impact of different environmental parameters on the fluctuation pattern of sensor data.

[0071] In another embodiment, the target parameter may further include a sensor type and a target object type. The sensor type may include at least one of a radar, a wall sensor, an image sensor, and an IMU sensor. That is, the sensor simulation data processing method of the present application may be used to implement simulation data processing of the above sensors.

[0072] The target object is the current sensor collection object, which is related to the application environment of the robot. For different robots, the setting of the target object can be different. For example, for a sweeping robot, the target object corresponding to it is an object that often appears in an indoor environment, such as a sofa, a tile, a mirror, a cabinet, etc. By using the simulation data processing method of the sensor of the present application, the mapping relationship between the target parameters of different target objects and the sensor data can be obtained.

[0073] In one embodiment, the target parameters include: sensor motion parameters, a relative distance between the sensor and the target object, and environmental parameters.

[0074] In one embodiment, the sensor moves through a sensor motion platform, wherein the sensor motion platform is a three-axis mobile platform in order to simulate the movement of a robot, and the sensor motion parameters may include the motion trajectory and motion speed of the sensor motion platform.

[0075] In another embodiment, Figure 4 As shown, controlling the real sensor to collect data of the target object with the target parameters to obtain the corresponding real sensor data includes the following steps:

[0076] Step 402: Control the environment adjustment device to provide a collection site environment based on environmental parameters.

[0077] Specifically, the environment adjustment device can be used to adjust the environment parameters, including humidity, temperature, light, etc. When the environment parameters change, the environment adjustment device can be controlled to adjust the environment parameters to provide a changing collection site environment.

[0078] Step 404 , controlling the object loading platform and the sensor motion platform to maintain a relative distance, and the target object is placed on the object loading platform.

[0079] Specifically, by setting the target object on the object loading platform, during adjustment, according to the relative distance between the sensor and the target object, the object loading platform is controlled to move to the corresponding position, so that the distance between the object loading platform and the sensor motion platform is a relative distance. Through the distance between the object loading platform and the sensor motion platform, the sensor can collect sensor data at different positions of the target object. It is understandable that if the target parameter that changes once does not include a change in distance, the object loading platform and the sensor motion platform do not need to be adjusted during this adjustment process. It is understandable that when the target object for the test needs to be replaced, the target object loaded on the object loading platform is replaced. For example, replace white tiles with black tiles.

[0080] Step 406 , controlling the sensor motion platform to move along the motion trajectory and at the motion speed, and controlling the real sensor to collect data on the target object during the motion process, to obtain corresponding real sensor data.

[0081] Under the set environmental parameters and distance, the sensor motion platform is controlled to move with the set motion trajectory and motion speed. At the same time, the real sensor collects data of the target object during the motion process to obtain the real sensor data corresponding to the target parameters. Through the motion trajectory and motion speed of the sensor motion platform, various movements of the robot can be simulated, such as forward, rotation, backward, and movement in all directions.

[0082] It should be noted that in the above steps, not all target parameters need to be adjusted during the sensor data collection process. When the target parameters change, only the changed target parameters need to be adjusted. For example, when the environmental parameters change, there is no need to adjust the distance, the motion trajectory and the motion speed of the sensor motion platform.

[0083] In this embodiment, the target parameters can be changed based on the sensor motion platform, the object loading platform and the environment adjustment device, thereby collecting rich sensor data.

[0084] In another embodiment, the target parameters also include acquisition parameters of the sensor. Specifically, the acquisition parameters of the sensor are calculated based on the effective real sensor data. The sensor acquisition parameters refer to the parameters used by the sensor when collecting data. For example, the focal length of the image sensor, the incident angle of the laser radar light, etc. By taking the sensor acquisition parameters as one of the dimensions of the target parameters, multiple acquisition parameters can be considered to explore the impact of different sensor acquisition parameters on the fluctuation law of sensor data.

[0085] In another embodiment, the simulation data processing method of the sensor further includes:

[0086] In response to the target object replacement instruction, the real sensor of the control robot is executed to process the target object with different target parameters and subsequent steps to obtain the mapping relationship between the target parameters and sensor data of the real sensor for different target objects.

[0087] Specifically, after completing the sensor data collection of a target object, the target object of the object loading platform can be replaced. For example, the material of the sofa can be replaced with the material of the tile, or the material of the white tile can be updated with the material of the black tile, and the target parameter-collection-change target parameter-collection cycle process is repeated to collect the real sensor data of different objects under different target parameters, and then fit the mapping relationship between the target parameters and sensor data of the target object, so as to obtain the mapping relationship between the target parameters and the real sensor data of different objects. For example, the mapping relationship between the target parameters and the real sensor data of white tiles and the mapping relationship between the target parameters and the real sensor data of black tiles are obtained.

[0088] In another embodiment, based on the effective real sensor data corresponding to different target parameters, a mapping relationship between the target parameters and sensor data of the real sensor to the target object is fitted, including: based on the effective real sensor data corresponding to different target parameters, a noise prediction model of the target object is trained; the noise prediction model is used to predict the noise to the target object according to the simulation parameters of the robot's simulation sensor in the simulation environment, and generate simulated sensor data containing noise according to the noise and sensor data of the simulated sensor.

[0089] Specifically, based on the effective real sensor data corresponding to different target parameters, a machine learning method can be used to train a noise prediction model to characterize the mapping relationship between the sensor's target parameters and noise data for the target object.

[0090] The noise prediction model can be deployed in the simulator. The simulated sensor collects sensor data from the target object under the simulation parameters. The simulation parameters are input into the noise prediction model. The noise to the target object is predicted based on the simulation parameters of the robot's simulated sensor in the simulation environment. Based on the noise and the sensor data of the simulated sensor, simulated sensor data containing noise is generated. Thus, the ideal values ​​returned by the sensors on the sweeper in the simulation environment are converted into sensor data that is close to the operating status of the sweeper in the real environment, and the role of simulation in algorithm development is played more efficiently.

[0091] In one embodiment, a method for processing simulation data of a sensor may include the following three stages:

[0092] Phase 1: Sensor data collection phase.

[0093] In the sensor data collection stage, the real sensor of the robot is controlled to collect data on the target object with different target parameters to obtain real sensor data corresponding to different target parameters; the valid data related to the target object in the real sensor data is screened to obtain valid real sensor data corresponding to different target parameters.

[0094] In the sensor data collection stage, the target parameters can be changed to obtain real sensor data corresponding to different target parameters.

[0095] In the sensor data collection stage, the target object may be changed to obtain real sensor data corresponding to different target parameters of different target objects.

[0096] The second stage: model training stage.

[0097] In the model training stage, the noise prediction model of the target object is trained based on the effective real sensor data corresponding to different target parameters.

[0098] It is understandable that, for real sensor data corresponding to different target parameters of different target objects, noise prediction models corresponding to different objects can be trained.

[0099] The noise prediction model is used to predict the noise to the target object according to the simulation parameters of the robot's simulation sensor in the simulation environment, and to generate simulated sensor data containing noise according to the noise and the sensor data in the simulated sensor.

[0100] In the model training phase, the noise model can be verified using real sensor data. The simulated parameter data corresponding to the same target parameters are compared with the real sensor data. If the error is less than the preset standard, the model verification is passed.

[0101] The third stage: simulation stage.

[0102] In the simulation stage, simulation parameters of the simulated sensor and sensor data of the simulated sensor are obtained, the simulation parameters are input into the noise prediction model to obtain noise data, the noise data and the data of the simulated sensor are superimposed to generate simulated sensor data containing noise.

[0103] The above-mentioned sensor simulation data processing method mines the fluctuation law of the sensor in the real environment to obtain the mapping relationship between the target parameter and the sensor data, and applies it to the sensor simulation to increase the reliability of the simulation results. In this way, more problems in the real machine are exposed in the simulation, and the efficiency of the simulation platform for algorithm iteration is improved.

[0104] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0105] Based on the same inventive concept, the embodiment of the present application also provides a sensor simulation data processing device for implementing the sensor simulation data processing method involved above. The implementation solutions provided by each device to solve the problem are similar to the implementation solutions recorded in the above method, so the specific limitations in the embodiments of each device provided below can refer to the limitations of the method above, and will not be repeated here.

[0106] In one embodiment, a sensor simulation data processing device is provided, such as Figure 5 As shown, including:

[0107] The acquisition module 502 is used to control the real sensor of the robot to collect data of the target object with different target parameters to obtain the real sensor data corresponding to the different target parameters;

[0108] A screening module 504 is used to screen the valid data related to the target object in the real sensor data to obtain the valid real sensor data corresponding to different target parameters;

[0109] The training module 506 is used to fit the mapping relationship between the target parameters and sensor data of the target object based on the effective real sensor data corresponding to different target parameters; the mapping relationship is used to take the simulation parameters of the robot's simulated sensor in the simulation environment as input, and output the simulated sensor data of the target object including noise.

[0110] In another embodiment, the acquisition module includes:

[0111] The parameter acquisition module is used to obtain the target parameters of the robot's real sensors.

[0112] The control module is used to control the real sensor to collect data on the target object with target parameters to obtain corresponding real sensor data.

[0113] Parameter change module, used to change target parameters.

[0114] In another embodiment, the target parameters include: sensor motion parameters, a relative distance between the sensor and the target object, and environmental parameters; the sensor motion parameters include a motion trajectory and a motion speed of the sensor motion platform.

[0115] The control module is used to control the environment adjustment device to provide a collection site environment based on environmental parameters; control the object loading platform to maintain a distance from the sensor motion platform, and the target object is set on the object loading platform; control the sensor motion platform to move at a motion trajectory and speed, and control the real sensor to collect data on the target object during the movement to obtain corresponding real sensor data.

[0116] In another embodiment, the acquisition module is further used to control the real sensor of the robot to collect data on the target object with different target parameters in response to the target object change instruction.

[0117] In another embodiment, a training module is used to train a noise prediction model of a target object based on valid real sensor data corresponding to different target parameters; the noise prediction model is used to predict the noise of the target object according to the simulation parameters of the robot's simulated sensor in a simulated environment, and to generate simulated sensor data containing noise according to the noise and sensor data of the simulated sensor.

[0118] Each module in the simulation data processing device of the sensor can be implemented in whole or in part by software, hardware and a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0119] In one embodiment, a computer device is provided, whose internal structure diagram can be as follows: Figure 6As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a simulation data processing method of a sensor is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0120] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0121] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the methods of the above embodiments when executing the computer program.

[0122] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the methods of the above embodiments are implemented.

[0123] In one embodiment, a computer program product is provided, including a computer program, which implements the steps of the methods of the above embodiments when executed by a processor.

[0124] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0125] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0126] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for processing simulation data of a sensor, It is characterized in that The method comprises: The real sensor of the robot is controlled to collect data of the target object with different target parameters, and the real sensor data corresponding to the different target parameters are obtained; Filtering valid data related to the target object in the real sensor data to obtain valid real sensor data corresponding to the different target parameters; Based on the effective real sensor data corresponding to the different target parameters, a mapping relationship between the target parameters and the sensor data of the target object is fitted; the mapping relationship is used to take the simulation parameters of the simulated sensor of the robot in the simulation environment as input, and output the simulated sensor data including noise for the target object.

2. The method according to claim 1, It is characterized in that The real sensor controlling the robot processes the target object with different target parameters to obtain real sensor data corresponding to the different target parameters, including: Get the target parameters of the robot's real sensors; Controlling the real sensor to collect data on the target object with the target parameter to obtain corresponding real sensor data; The target parameter is changed, and the step of obtaining the target parameter of the real sensor of the robot is returned to obtain the real sensor data corresponding to different target parameters.

3. The method according to claim 1 or 2, It is characterized in that The target parameter includes at least one of a sensor motion parameter, a relative distance between the sensor and the target object, a sensor acquisition parameter, and an environmental parameter.

4. The method according to claim 3, It is characterized in that The target parameters include: sensor motion parameters, the relative distance between the sensor and the target object, and environmental parameters; the sensor motion parameters include the motion trajectory and motion speed of the sensor motion platform; The controlling the real sensor to collect data on the target object with the target parameter to obtain corresponding real sensor data includes: Controlling the environment adjustment device to provide a collection site environment based on the environmental parameters; Controlling the object loading platform and the sensor motion platform to maintain the relative distance, wherein the target object is disposed on the object loading platform; The sensor motion platform is controlled to move along the motion trajectory and at the motion speed, and the real sensor is controlled to collect data on the target object during the motion to obtain corresponding real sensor data.

5. The method according to claim 1, It is characterized in that The method further comprises: In response to the target object replacement instruction, the real sensor of the control robot is executed to collect data on the target object with different target parameters and subsequent steps to obtain a mapping relationship between the target parameters and sensor data of the real sensor for different target objects.

6. The method according to claim 1 or 5, It is characterized in that Based on the effective real sensor data corresponding to the different target parameters, a mapping relationship between the target parameter and the sensor data of the real sensor and the target object is obtained by fitting, including: Based on the effective real sensor data corresponding to the different target parameters, a noise prediction model of the target object is trained; the noise prediction model is used to predict the noise of the target object according to the simulation parameters of the robot's simulation sensor in the simulation environment, and generate simulated sensor data containing noise according to the noise and the sensor data of the simulated sensor.

7. A sensor simulation data processing device, It is characterized in that include: An acquisition module is used to control the real sensor of the robot to collect data of the target object with different target parameters to obtain the real sensor data corresponding to the different target parameters; A screening module, used for screening the valid data related to the target object in the real sensor data, and obtaining the valid real sensor data corresponding to the different target parameters; A prediction module, configured to obtain a mapping relationship between the target parameters and the sensor data of the target object by fitting based on the effective real sensor data corresponding to the different target parameters; The mapping relationship is used to take the simulation parameters of the simulation sensor of the robot in the simulation environment as input, and output the simulation sensor data including noise for the target object.

8. A sensor simulation data processing system, It is characterized in that include: A data acquisition device, used to control the real sensor of the robot to collect data on the target object with different target parameters, and obtain the real sensor data corresponding to the different target parameters; A host computer, connected to the data acquisition device, for screening the valid data related to the target object in the real sensor data, and obtaining the valid real sensor data corresponding to the different target parameters; Based on the effective real sensor data corresponding to the different target parameters, a mapping relationship between the target parameters and the sensor data of the target object is obtained by fitting; The mapping relationship is used to take the simulation parameters of the simulation sensor of the robot in the simulation environment as input, and output the simulation sensor data including noise for the target object.

9. The system according to claim 8, It is characterized in that The data acquisition device comprises: Sensor motion platform, used to load real sensors; The object loading platform is used to load the target object; the sensor motion platform is arranged opposite to the object loading platform; Environmental adjustment device, used for adjusting environmental parameters; a controller, electrically connected to the sensor motion platform, the object loading platform and the environment adjustment device, and configured to control at least one of the sensor motion platform, the object loading platform and the environment adjustment device to adjust according to the target parameter; The host computer is directly or indirectly connected to the real sensor to obtain real sensor data.

10. The system according to claim 9, It is characterized in that The target parameters include: sensor motion parameters, the relative distance between the sensor and the target object, and environmental parameters; the sensor motion parameters include the motion trajectory and motion speed of the sensor motion platform; the controller is used to control the environmental adjustment device to provide a collection site environment based on the environmental parameters; control the object loading platform and the sensor motion platform to maintain the relative distance, and the target object is set on the object loading platform; control the sensor motion platform to move with the motion trajectory and the motion speed, and control the real sensor to collect data on the target object during the movement to obtain corresponding real sensor data.

Images