Three-axis robot control method and device

Abstract

The present disclosure provides a three-axis robot control method and device, belonging to the technical field of robot control. The method includes: identifying a target object by identifying a plurality of first objects based on a target recognition algorithm; controlling the three-axis robot to perform line-of-sight tracking on the target object based on a tracking algorithm, where the line-of-sight tracking is the tracking of the target object by an in-built camera of the three-axis robot; and controlling the three-axis robot to perform motion tracking on the target object based on a path planning algorithm, where the motion tracking is the tracking of the target object by the movement of the three-axis robot. The three-axis robot control method and device provided by the present disclosure can improve the accuracy and reliability of target tracking, and further improve the accuracy and reliability of robot control.

Description

Technical Field

[0001] This disclosure belongs to the technical field of robot control, and more specifically, relates to a three-axis robot control method and device. Background Art

[0002] In the field of modern automation control and monitoring, the technology of target recognition and tracking is crucial. Three-axis robots undertake the tasks of identifying and tracking target objects in many scenarios, such as an automatic recognition type three-axis robot accurately locking a drone target, and a three-axis robot carried by a high-altitude drone automatically identifying and tracking a ground target.

[0003] Due to the interference of the external environment, the traditional target tracking method has low accuracy and reliability in tracking the target, so it is impossible to control the three-axis robot to effectively track the target. Therefore, there is an urgent need for a stable and reliable three-axis robot control method. Summary of the Invention

[0004] The purpose of this disclosure is to provide a three-axis robot control method and device to improve the accuracy and reliability of target tracking, and further improve the accuracy and reliability of robot control.

[0005] In the first aspect of the embodiments of this disclosure, a three-axis robot control method is provided, including:

[0006] Identifying target objects from multiple first objects based on a target recognition algorithm;

[0007] Controlling the three-axis robot to perform line-of-sight tracking on the target objects based on a tracking algorithm, where the line-of-sight tracking is the tracking of the target objects by the built-in camera of the three-axis robot;

[0008] Controlling the three-axis robot to perform motion tracking on the target objects based on a path planning algorithm, where the motion tracking is the tracking of the target objects by the movement of the three-axis robot.

[0009] In the second aspect of the embodiments of this disclosure, a three-axis robot control device is provided, including:

[0010] A target recognition module for identifying target objects from multiple first objects based on a target recognition algorithm;

[0011] A line-of-sight tracking module for controlling the three-axis robot to perform line-of-sight tracking on the target objects based on a tracking algorithm, where the line-of-sight tracking is the tracking of the target objects by the built-in camera of the three-axis robot;

[0012] A motion tracking module for controlling the three-axis robot to perform motion tracking on the target objects based on a path planning algorithm, where the motion tracking is the tracking of the target objects by the movement of the three-axis robot.

[0013] In a third aspect of the embodiments of the present disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the above-mentioned three-axis robot control method are implemented.

[0014] In a fourth aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned three-axis robot control method are implemented.

[0015] The beneficial effects of the three-axis robot control method and device provided by the embodiments of the present disclosure are as follows:

[0016] Through the target recognition algorithm, the present disclosure can quickly identify the target object from multiple first objects, and in combination with the tracking algorithm, the three-axis robot can perform line-of-sight tracking on the target object in real time to ensure that the robot can stably track the target. Based on the path planning algorithm, the three-axis robot can plan the optimal path from the current position to the position of the target object to ensure that the robot can efficiently reach the position during tracking, improving the accuracy and reliability of target tracking, and further improving the accuracy and reliability of robot control. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the following drawings are only some embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a schematic flowchart of the three-axis robot control method provided by an embodiment of the present disclosure;

[0019] Figure 2 It is a structural block diagram of the three-axis robot control device provided by an embodiment of the present disclosure;

[0020] Figure 3 It is a schematic block diagram of the electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] In the following description, specific details such as specific system structures and technologies are presented for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of the present disclosure. However, those skilled in the art should clearly understand that the present disclosure can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from interfering with the description of the present disclosure.

[0022] To make the objectives, technical solutions, and advantages of the present disclosure clearer, the following will be described through specific embodiments in conjunction with the accompanying drawings.

[0023] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of a three-axis robot control method provided by an embodiment of the present disclosure. The method includes:

[0024] S101: Identify target objects by identifying multiple first objects based on a target recognition algorithm.

[0025] In this embodiment, the target recognition algorithm can be a feature extraction algorithm, such as Scale-Invariant Feature Transform (SIFT), Histogram of Oriented Gradients (HOG), etc., or a trained Support Vector Machine (SVM) model or neural network model, etc. The purpose is to determine the object to be tracked from multiple objects.

[0026] The first object refers to an object that appears within the effective detection area of the robot. The target object is a preset object that needs to be tracked. For example, in our application scenario, the robot is used to detect and track an unmanned aerial vehicle (UAV) in the air on the ground. If a flock of birds, an airplane, and a UAV appear within the effective detection area of the robot, then the flock of birds, the airplane, and the UAV are all first objects. A trained neural network model can be used to identify the target object, i.e., the UAV, from the images collected by the robot.

[0027] S102: Control the three-axis robot to perform line-of-sight tracking on the target object based on a tracking algorithm. The line-of-sight tracking is the tracking of the target object by the built-in camera of the three-axis robot.

[0028] In this embodiment, the tracking algorithm can be a particle filter algorithm. The three-axis robot has three movable axes and can move and rotate in different directions in space, so as to perform tasks such as operating and tracking the target object.

[0029] The target object is the object that needs to be tracked. The line-of-sight tracking means to continuously track the target in real time by rotating the built-in camera of the three-axis robot to prevent the loss of the target and cause the tracking to fail. For example, when it is detected that the target object appears in the right edge area of the image, the camera is controlled to rotate to the right, and the same applies to other directions. The built-in camera can be a camera device installed inside or on the body of the three-axis robot, which is used to capture images or videos of the target object and provide information about the target object for the tracking algorithm.

[0030] S103: Control the three-axis robot to perform motion tracking on the target object based on a path planning algorithm. The motion tracking is the tracking of the target object by the movement of the three-axis robot.

[0031] In this embodiment, the path planning algorithm can be Dijkstra's algorithm or A* algorithm. According to the current position of the robot, the position of the target object, and the distribution of surrounding environmental obstacles, an optimal path from the current position of the robot to the position of the target object is calculated. The control system sends control instructions to each axis of the three-axis robot according to the path planning result to adjust the movement of the axis, so that the robot moves along the planned path.

[0032] That is, gaze tracking is to control the movement of the camera built in the robot, while motion tracking is to control the movement of the robot.

[0033] It can be concluded from the above that the present disclosure can quickly identify the target object from multiple first objects through the target recognition algorithm, and in combination with the tracking algorithm, the three-axis robot can perform real-time gaze tracking on the target object to ensure that the robot can stably track the target. Based on the path planning algorithm, the three-axis robot can plan an optimal path from the current position to the position of the target object to ensure that the robot can efficiently reach the position during tracking, improving the accuracy and reliability of target tracking, and further improving the accuracy and reliability of robot control.

[0034] In an embodiment of the present disclosure, the method for controlling a three-axis robot further includes:

[0035] Determining the position error, speed error, and basic motion characteristics of the target object based on the radar.

[0036] The tracking algorithm is a particle filter algorithm;

[0037] Controlling the three-axis robot to perform gaze tracking on the target object based on the tracking algorithm, including:

[0038] Determining the number of particles based on the basic motion characteristics;

[0039] Determining the initial weight of each particle based on the number of particles;

[0040] Initializing the particle filter algorithm based on the number of particles, the initial weight of each particle, the position error, and the speed error;

[0041] Determining the motion model based on the basic motion characteristics, and determining the predicted state of the target object based on the motion model and the initialized particles;

[0042] Controlling the three-axis robot to perform gaze tracking on the target object based on the predicted state.

[0043] In this embodiment, the purpose of the radar is to first perform preliminary monitoring on the target object to make the subsequent particle filter algorithm more accurate and reliable. The radar can be the radar of the three-axis robot itself, or it can not be set in the three-axis robot, and the data detected by relevant equipment is sent to the three-axis robot.

[0044] Position error refers to the distance between the possible position of the target object and the predicted position, and velocity error refers to the difference between the possible velocity of the target and the predicted velocity. For example, in the application scenario where a three-axis robot locks onto and tracks a drone target, due to the fast movement speed, strong maneuverability, and complex state changes of the drone in the air, the drone can move in a large airspace. Compared with ground targets, its position error and velocity error are relatively large.

[0045] Basic motion characteristics refer to some basic descriptions of the motion state of the target object, such as whether the target object is moving at a constant speed or accelerating.

[0046] The position error, velocity error, and basic motion characteristics of the target object can be preliminarily detected by radar in advance to determine the speed and the rate of change of the speed of the target object. For example, it can be obtained through a preset first mapping table, such as Table 1.

[0047]

[0048] It should be noted that Table 1 is only part of the data of the first mapping table. In addition, there are other value ranges, and the position error, velocity error, etc. can be limited according to a wider range of speeds or range of speed change rates.

[0049] In this embodiment, corresponding conditions can also be preset in advance. For example, in response to the basic moving speed of the target object being greater than the first speed threshold, the first speed error is used as the velocity error, and the first position error is used as the position error;

[0050] In response to the basic moving speed of the target object being less than or equal to the first speed threshold, the second speed error is used as the velocity error, and the second position error is used as the position error.

[0051] The basic moving speed is the moving speed of the target object detected by radar for the target object. The position error and velocity error can be determined accordingly. The first speed threshold can be preset according to experience. The first speed error is greater than the second speed error, and the first position error is greater than the second position error. The first speed error, the first position error, the second speed error, and the second position error can all be determined according to the actual situation.

[0052] In this embodiment, determining the number of particles based on the basic motion characteristics includes: in response to the basic motion characteristic being uniform motion, determining the number of particles as the first number of particles; in response to the basic motion characteristic being non-uniform motion, determining the number of particles as the second number of particles, where the first number of particles is less than the second number of particles. This is because if the motion of the target object is relatively complex and uncertain, more particles are needed to more accurately cover the possible state space of the target; conversely, if the motion of the target object is relatively simple and stable, the number of particles can be appropriately reduced. Therefore, when the basic motion characteristic is uniform motion, a smaller number can be set, and when the basic motion characteristic is non-uniform motion, a larger number can be set. The first number of particles and the second number of particles are preset values, which can be determined by the number of particles when solving similar problems in this field or set according to experience.

[0053] When initializing the particle filter algorithm, an initial weight is assigned to each particle, and the initial weights of each particle are equal because at this time, the target object has not been tracked yet, and in the absence of other known information, the weights of each particle should be the same, that is, the initial weight = 1 / the number of particles. For example, if the number of particles is 500, the initial weight of each particle is 1 / 500.

[0054] Comprehensively utilize the previously determined number of particles, the initial weight of each particle, and the position error and velocity error obtained by the radar to set the initial state of the particle filter algorithm, and prepare for subsequent use of this algorithm for target state estimation and tracking.

[0055] Determining the motion model based on the basic motion characteristics means selecting or constructing a suitable motion model through the basic motion characteristics of the target object previously determined by the radar. For example, in response to the basic motion characteristic being uniform motion, determining the uniform linear motion model as the motion model; in response to the basic motion characteristic being non-uniform motion, determining the uniformly accelerated motion model as the motion model.

[0056] Using the determined motion model, combined with the initialized particles, through the calculation and reasoning of the particle filter algorithm, predict the state of the target object at the current moment or future moments, such as position and velocity (i.e., the predicted state). Based on the predicted state of the target object obtained by the particle filter algorithm, send a control command to the three-axis robot so that the built-in camera of the robot can always be aimed at the target object, realizing line-of-sight tracking of the target object and preventing tracking failure.

[0057] In this embodiment, the present disclosure determines the position error, velocity error, and basic motion characteristics of the target object by introducing a radar, providing more accurate and comprehensive initial information for the subsequent particle filter algorithm, improving the convergence speed and tracking accuracy of the tracking algorithm. Especially in the case where the motion state of the target object is complex and variable, the application of the particle filter algorithm, combined with the information provided by the radar, enables the three-axis robot to effectively cope with the uncertainty of the target object in a dynamic environment, enhancing the robustness of the present disclosure. This embodiment dynamically adjusts the number of particles according to the basic motion characteristics, ensuring the accuracy of tracking, reducing the computational complexity, improving the accuracy and reliability of target tracking, and further enhancing the accuracy and reliability of robot control.

[0058] In an embodiment of the present disclosure, the three-axis robot control method further includes:

[0059] In response to the time for controlling the three-axis robot to perform line-of-sight tracking on the target object based on the predicted state exceeding a first duration;

[0060] Determine the observation noise based on the noise characteristics, and determine the observation model based on the observation noise;

[0061] Update the initial weight of each particle based on the observation model to obtain the first weight of each particle;

[0062] In response to the first weight satisfying the resampling condition, resample each particle to obtain particles with a second number of particles, and determine the second predicted state of the target object based on the particles with the second number of particles and the motion model;

[0063] Control the three-axis robot to perform line-of-sight tracking on the target object based on the second predicted state.

[0064] In this embodiment, it is considered that the prediction step infers the state at the current moment based on the state and motion model at the previous moment. However, the motion model itself is an approximation of the actual motion and has a certain degree of inaccuracy, and over time, these errors will accumulate. Relying solely on the predicted state, the error may become larger and larger, resulting in the tracking of the target gradually deviating from the true trajectory.

[0065] Therefore, after the time for controlling the three-axis robot to perform line-of-sight tracking on the target object based on the predicted state exceeds the first duration, the weights of each particle should be updated and the position should be re-predicted to better track the target object. The first duration can be determined according to the actual situation during the experiment. For example, if it is found that the target cannot be effectively tracked after 10 seconds, the first duration can be set to 10 seconds.

[0066] The noise characteristics refer to the noise caused by the occlusion of the target object, direct sunlight in the environment, etc. during the tracking process.

[0067] Determine the observation noise based on the noise characteristics, and determine the observation model based on the observation noise, including: in response to the target object being occluded and / or the environmental light shining directly, determine the observation noise based on the first observation noise covariance matrix;

[0068] In response to the target object not being occluded and the environmental light not shining directly, determine the observation noise based on the second observation noise covariance matrix.

[0069] The first observation noise covariance matrix can be and the second observation noise covariance matrix can be The meaning of 50 in the observation noise covariance matrix is the noise variance for position observation, which is set to 50 square meters in the x, y, and z directions. The purpose is to prevent tracking failure and reduce errors.

[0070] The observation model can be where is the observation value at time , is the observation matrix, represents the state vector of the target object at time , is the observation noise vector, and or , represents time.

[0071] is obtained through the motion model in the foregoing part. For example, the above motion model is a uniformly accelerated motion, and its motion model is where represents the state vector of the target object at time , represents the state vector of the target object at time , is the velocity vector at time , is the acceleration vector, is the time interval. The above motion model calculates its state at time, which can include the position and velocity information of the target object, and can be expressed as , that is, the position coordinates and the velocity information in three directions.

[0072] The element values of the observation matrix depend on the specific observation method and physical model, and can be in this application, and can also be set based on experience and the reference matrix when solving similar problems.

[0073] Subsequently, the observation value Calculate the importance weight of each particle. The calculation of the weight is based on the likelihood of the observation value in the state of the particle, that is, update the initial weight of each particle based on the observation model to obtain the first weight of each particle, including:

[0074] Determine the observation likelihood of each particle based on the observation model;

[0075] Introduce the time decay factor into the weight update formula to obtain the first weight update formula;

[0076] Update the initial weight of each particle based on the observation likelihood of each particle and the first weight update formula to obtain the first weight of each particle.

[0077] Specifically, the calculation of the weight is based on the likelihood probability of the observation value in the state of the particle, that is , where , represents the state vector of the target object at time for the th particle, is the velocity vector of the th particle at time , is the acceleration vector of the th particle at time , is the observation value at time .

[0078] The weight update formula can be , where is the observation dimension, which is 2 in this application.

[0079] Among them, , represents the transpose, and the denominator is the normalization constant, which is used to ensure that the integral of the probability density function over the entire observation space is equal to 1. The exponential part in the numerator reflects the difference degree between the observation value and the particle prediction state . is the residual vector between the observation value and the predicted observation value. Weigh the residuals. The inverse of the covariance matrix makes the influence of the residuals relatively small in the direction with large noise and relatively large in the direction with small noise. Overall, the larger the value of this exponential term, the closer the observation value is to the particle prediction state, and the higher the particle weight. Through such a calculation method, the particle filter algorithm can update the particle weights according to the observation data, and then screen out the particles that are more in line with the true state of the target.

[0080] In this embodiment, the time decay factor is also considered. The state of the target object changes over time. For example, when tracking a moving target object, its motion pattern, speed, direction, etc. may change. The descriptive ability of the observation data and particle weights at an earlier time for the target state at the current time will gradually weaken. Introducing the time decay factor can make the algorithm pay more attention to recent observations and state information, enabling the weight update to better adapt to the dynamic changes of the target state and improving the tracking accuracy and adaptability of the target. For example, if the target object was originally moving in a straight line at a constant speed and suddenly starts to turn, the time decay factor can enable the algorithm to adjust the particle weights more quickly according to the new observations without overly relying on the weight information during the previous uniform motion.

[0081] The update formula for the weight after introducing the time decay factor is the first weight update formula, which can be , denotes the th particle at time The weight value after introducing the time decay factor, that is, the first weight, denotes the time decay factor, which can be determined according to experience or experimental process data. According to the first weight update formula and the observation likelihood of each particle, the initial weight of each particle can be updated to obtain the first weight of each particle.

[0082] It should be noted that the first weight should be normalized to prevent the sum of the first weights from not being 1. After the above description, the first weight of each particle can be obtained. When the first weight meets the resampling condition, resampling is performed. Whether the resampling condition is met can be judged in the following way:

[0083] Determine the number of effective particles based on the first weight of each particle.

[0084] In response to the number of effective particles being less than the effective value, the first weight meets the resampling condition.

[0085] In response to the number of effective particles being greater than or equal to the effective value, the first weight does not meet the resampling condition.

[0086] The number of effective particles can be determined according to the effective particle number calculation formula. For example, the effective particle number calculation formula is , and the effective value can be determined according to the actual scenario.

[0087] As can be seen from the above, by introducing an observation model and dynamically determining the observation noise according to the noise characteristics, the present disclosure can more accurately describe the observation process of the target object, reducing the tracking deviation caused by observation errors. In this embodiment, by considering the time decay factor in the weight update mechanism, the present disclosure can pay more attention to recent observation data, improving the adaptability to the dynamic changes of the target object state, thereby improving the accuracy and stability of tracking, and further enhancing the accuracy and reliability of robot control. In this embodiment, by setting an effective particle number threshold and triggering resampling only when the weight distribution is extremely uneven, unnecessary computational overhead is avoided, improving the efficiency of the present disclosure.

[0088] In an embodiment of the present disclosure, the three-axis robot control method further includes:

[0089] In response to the change rate of the second predicted state for a continuous first number being less than the first change rate, increase the time decay factor based on the first time step;

[0090] In response to the change rate of the second predicted state for a continuous first number being greater than the second change rate, decrease the time decay factor based on the second time step.

[0091] In this embodiment, when the change rate of the second predicted state for a continuous first number is less than the first change rate, it indicates that the state change of the target object is relatively slow and stable. At this time, increase the time decay factor based on the first time step. Increasing the time decay factor will increase the proportion of the historical weight in the current weight calculation. Because the current target state is relatively stable and the historical information has high reference value, this can make the algorithm more smoothly track the target state and reduce the drastic weight fluctuations caused by single observation errors.

[0092] If the change rate of the second predicted state for a continuous first number is greater than the second change rate, it indicates that the state change of the target object is relatively fast. At this time, decrease the time decay factor based on the second time step. Decreasing the time decay factor will attenuate the influence of the historical weight and pay more attention to the current observation data, enabling faster tracking of the sudden change of the target state and timely adjusting the estimation of the target object state. The first change rate and the second change rate are empirically preset values.

[0093] The change rate of the second predicted state can be determined by the similarity of the second predicted states in two consecutive frames. For example, it can be calculated by means such as cosine similarity. The higher the similarity, the lower the change rate. The first time step and the second time step can be determined according to the change rate difference. For example, the first time step , the second time step , where is the initial value of the first time step and the second time step, are linear coefficients for adjusting the first time step and the second time step, respectively, which can be determined according to experiments. is the actual change rate. is the first change rate. is the second change rate.

[0094] It can be concluded from the above that by dynamically adjusting the time decay factor, the present disclosure can adjust the dependence on historical information and current observation data according to the speed of change of the target object state, so that when the three-axis robot faces target objects with different dynamic characteristics, it can maintain stable tracking performance and improve the adaptability and robustness of tracking.

[0095] In an embodiment of the present disclosure, determining the predicted state of the target object based on the motion model and the initialized particles includes:

[0096] Determining the noise-free predicted state of the target object based on the motion model;

[0097] In response to the environmental characteristics satisfying the first environmental condition, taking the noise-free predicted state as the predicted state;

[0098] In response to the environmental characteristics not satisfying the first environmental condition, determining Gaussian noise based on the environmental characteristics, and determining the predicted state based on the Gaussian noise and the noise-free predicted state; the first environmental condition includes that the environmental wind speed is less than the first wind speed.

[0099] In this embodiment, during the prediction process, the uncertainty of object motion is considered, such as the influence of factors such as wind direction and air flow on the target motion. This uncertainty can be simulated by adding random noise to the particles. For example, Gaussian noise is added to the velocity component. When the wind speed is less than the first wind speed, it is considered that the influence on the target object is small, and at this time, adding noise is not considered. When the wind speed is greater than or equal to the first wind speed, it is considered that the influence on the target object is large, and at this time, Gaussian noise should be added.

[0100] Specifically, Gaussian noise can be added to the velocity component , that is , where , is the covariance matrix of the process noise.

[0101] The environmental characteristics include wind speed. The greater the wind speed, the larger the value range of the corresponding Gaussian noise, that is, the larger the value in the covariance matrix of the process noise. A mapping table can be preset to determine the value in the covariance matrix of the process noise according to the wind speed. For example, the covariance matrix of the process noise corresponding to a wind speed of 5 meters per second is:

[0102]

[0103] In the matrix, "100" represents the noise variance of the position component, and "50" represents the noise variance of the velocity components in the x, y, and z directions.

[0104] As can be seen from the above, by considering the influence of the environmental wind speed on the movement of the target object and dynamically adjusting the prediction model (whether to add Gaussian noise and its magnitude) according to the wind speed, the present disclosure can more accurately predict the future state of the target object, improve the accuracy and reliability of target tracking, and further improve the accuracy and reliability of robot control.

[0105] Corresponding to the three-axis robot control method in the above embodiment, Figure 2 is a structural block diagram of a three-axis robot control device provided by an embodiment of the present disclosure. For the sake of convenience of description, only the parts related to the embodiments of the present disclosure are shown. Refer to Figure 2 The three-axis robot control device 20 includes: a target recognition module 21, a line-of-sight tracking module 22, and a motion tracking module 23;

[0106] Among them, the target recognition module 21 is used to identify a plurality of first objects based on a target recognition algorithm to obtain a target object;

[0107] The line-of-sight tracking module 22 is used to control the three-axis robot to perform line-of-sight tracking on the target object based on a tracking algorithm. The line-of-sight tracking is the tracking of the target object by an in-built camera of the three-axis robot;

[0108] The motion tracking module 23 is used to control the three-axis robot to perform motion tracking on the target object based on a path planning algorithm. The motion tracking is the tracking of the target object by the movement of the three-axis robot.

[0109] In an embodiment of the present disclosure, the three-axis robot control device 20 further includes: a radar processing module, which is used to determine the position error, velocity error, and basic motion characteristics of the target object based on a radar.

[0110] In an embodiment of the present disclosure, the tracking algorithm is a particle filter algorithm;

[0111] The line-of-sight tracking module 22 is specifically used to determine the number of particles based on the basic motion characteristics;

[0112] Determine the initial weight of each particle based on the number of particles;

[0113] Initialize the particle filter algorithm based on the number of particles, the initial weight of each particle, the position error, and the velocity error;

[0114] Determine a motion model based on the basic motion characteristics, and determine the predicted state of the target object based on the motion model and the initialized particles;

[0115] The three-axis robot is controlled based on the predicted state to perform line-of-sight tracking on the target object.

[0116] In one embodiment of the present disclosure, the three-axis robot control device 20 further includes: a weight update module, configured to respond to the time for controlling the three-axis robot to perform line-of-sight tracking on the target object based on the predicted state exceeding a first duration;

[0117] Determine the observation noise based on the noise characteristics, and determine the observation model based on the observation noise;

[0118] Update the initial weight of each particle based on the observation model to obtain the first weight of each particle;

[0119] In response to the first weight satisfying the resampling condition, resample each particle to obtain particles with a second number of particles, and determine the second predicted state of the target object based on the particles with the second number of particles and the motion model; <http: / / www.example.com /

[0120] Control the three-axis robot to perform line-of-sight tracking on the target object based on the second predicted state.

[0121] In one embodiment of the present disclosure, the weight update module is specifically configured to determine the observation likelihood of each particle based on the observation model;

[0122] Introduce a time decay factor into the weight update formula to obtain a first weight update formula;

[0123] Update the initial weight of each particle based on the observation likelihood of each particle and the first weight update formula to obtain the first weight of each particle.

[0124] In one embodiment of the present disclosure, the three-axis robot control device 20 further includes: a time decay adjustment module, configured to respond to the change rate of the second predicted state of a continuous first number being less than a first change rate, and increase the time decay factor based on a first time step;

[0125] In response to the change rate of the second predicted state of a continuous first number being greater than a second change rate, reduce the time decay factor based on a second time step.

[0126] In one embodiment of the present disclosure, the line-of-sight tracking module 22 is specifically further configured to determine the noise-free predicted state of the target object based on the motion model;

[0127] In response to the environmental characteristics satisfying a first environmental condition, use the noise-free predicted state as the predicted state;

[0128] In response to the environmental characteristics not satisfying the first environmental condition, determine Gaussian noise based on the environmental characteristics, and determine the predicted state based on the Gaussian noise and the noise-free predicted state; the first environmental condition includes that the environmental wind speed is less than a first wind speed.

[0129] See Figure 3 , Figure 3 which is a schematic block diagram of an electronic device provided by an embodiment of the present disclosure. As shown in Figure 3 , the electronic device 300 in this embodiment may include: one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The above-mentioned processors 301, input devices 302, output devices 303, and memories 304 communicate with each other through a communication bus 305. The memory 304 is used to store computer programs, and the computer programs include program instructions. The processor 301 is used to execute the program instructions stored in the memory 304. Among them, the processor 301 is configured to call the program instructions to execute the functions of each module / unit in the above-mentioned device embodiments, for example Figure 2 the functions of the modules 21 to 23 shown

[0130] It should be understood that in the embodiments of the present disclosure, the so-called processor 301 may be a central processing unit (CPU), and this processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or this processor may also be any conventional processor, etc.

[0131] The input device 302 may include a touchpad, a fingerprint acquisition sensor (for acquiring the fingerprint information and the direction information of the fingerprint of the user), a microphone, etc., and the output device 303 may include a display (such as an LCD), a speaker, etc.

[0132] The memory 304 may include a read-only memory and a random access memory, and provide instructions and data to the processor 301. A part of the memory 304 may also include a non-volatile random access memory. For example, the memory 304 may also store information about the device type.

[0133] In specific implementation, the processors 301, input devices 302, and output devices 303 described in the embodiments of the present disclosure may implement the implementation manners described in the first and second embodiments of the three-axis robot control method provided by the embodiments of the present disclosure, and may also implement the implementation manner of the electronic device described in the embodiments of the present disclosure, which will not be elaborated here.

[0134] In another embodiment of the present disclosure, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program, and the computer program includes program instructions. When the program instructions are executed by a processor, all or part of the processes in the methods of the above embodiments are implemented. It can also be completed by instructing relevant hardware through the computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0135] The computer-readable storage medium can be the internal storage unit of the electronic device in any of the foregoing embodiments, such as the hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk equipped on the electronic device, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Further, the computer-readable storage medium can also include both the internal storage unit and the external storage device of the electronic device. The computer-readable storage medium is used to store the computer program and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store the data that has been output or will be output.

[0136] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present disclosure.

[0137] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described electronic devices and units can refer to the corresponding processes in the foregoing method embodiments and will not be described in detail here.

[0138] In several embodiments provided by the present application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed coupling, direct coupling, or communication connection to each other can be an indirect coupling or communication connection through some interfaces or units, and can also be in the form of electrical, mechanical, or other connections.

[0139] The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiments of the present disclosure.

[0140] In addition, in each embodiment of the present disclosure, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0141] The above are only the specific implementation manners of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present disclosure can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should all be covered within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be subject to the protection scope of the claims.

Claims

1. A three-axis robot control method, characterized in that, Including: Identifying a target object by identifying a plurality of first objects based on a target recognition algorithm; Determining a position error, a velocity error, and a basic motion feature of the target object based on a radar; Determining the number of particles based on the basic motion feature; Determining an initial weight of each particle based on the number of particles; Initializing a particle filter algorithm based on the number of particles, the initial weight of each particle, the position error, and the velocity error; Determining a motion model based on the basic motion feature, and determining a predicted state of the target object based on the motion model and the initialized particles; Controlling a three-axis robot to perform line-of-sight tracking on the target object based on the predicted state; the line-of-sight tracking is the tracking of the target object by an in-built camera of the three-axis robot; Controlling the three-axis robot to perform motion tracking on the target object based on a path planning algorithm, the motion tracking being the tracking of the target object by the movement of the three-axis robot; In response to the time for controlling the three-axis robot to perform line-of-sight tracking on the target object based on the predicted state exceeding a first duration; Determining an observation noise based on a noise feature, and determining an observation model based on the observation noise; Determining an observation likelihood of each particle based on the observation model; Introducing a time decay factor into a weight update formula to obtain a first weight update formula; Updating the initial weight of each particle based on the observation likelihood of each particle and the first weight update formula to obtain a first weight of each particle; In response to the first weight satisfying a resampling condition, resampling each particle to obtain particles with a second number of particles, and determining a second predicted state of the target object based on the particles with the second number of particles and the motion model; Controlling the three-axis robot to perform line-of-sight tracking on the target object based on the second predicted state; In response to the change rate of a continuous first number of second predicted states being less than a first change rate, increasing the time decay factor based on a first time step; In response to the change rate of a continuous first number of second predicted states being greater than a second change rate, decreasing the time decay factor based on a second time step.

2. The three-axis robot control method according to claim 1, characterized in that, The determining the predicted state of the target object based on the motion model and the initialized particles includes: Determining a noise-free predicted state of the target object based on the motion model; In response to the environmental feature satisfying a first environmental condition, using the noise-free predicted state as the predicted state; In response to the environmental feature not satisfying the first environmental condition, determining Gaussian noise based on the environmental feature, and determining the predicted state based on the Gaussian noise and the noise-free predicted state; the first environmental condition includes that the environmental wind speed is less than a first wind speed.

3. A three-axis robot control device, characterized in that, Including: A target recognition module for identifying a target object by identifying a plurality of first objects based on a target recognition algorithm; A radar processing module for determining a position error, a velocity error, and a basic motion feature of a target object based on a radar; A line-of-sight tracking module for determining the number of particles based on a basic motion feature; Determining an initial weight of each particle based on the number of particles; Initializing a particle filter algorithm based on the number of particles, the initial weight of each particle, the position error, and the velocity error; Determine a motion model based on basic motion features, and determine the predicted state of the target object based on the motion model and the initialized particles; Control the three-axis robot to perform line-of-sight tracking on the target object based on the predicted state; the line-of-sight tracking is the tracking of the target object by the built-in camera of the three-axis robot; A motion tracking module, configured to control the three-axis robot to perform motion tracking on the target object based on a path planning algorithm, where the motion tracking is the tracking of the target object by the movement of the three-axis robot; A weight update module, configured to respond to the time for controlling the three-axis robot to perform line-of-sight tracking on the target object based on the predicted state exceeding a first duration; Determine the observation noise based on the noise characteristics, and determine the observation model based on the observation noise; Determine the observation likelihood of each particle based on the observation model; Introduce a time decay factor into the weight update formula to obtain a first weight update formula; Update the initial weight of each particle based on the observation likelihood of each particle and the first weight update formula to obtain the first weight of each particle; In response to the first weight satisfying the resampling condition, resample each particle to obtain particles with a second number of particles, and determine the second predicted state of the target object based on the particles with the second number of particles and the motion model; Control the three-axis robot to perform line-of-sight tracking on the target object based on the second predicted state; A time decay adjustment module, configured to respond to the change rate of a continuous first number of second predicted states being less than a first change rate, and increase the time decay factor based on a first time step; In response to the change rate of a continuous first number of second predicted states being greater than a second change rate, reduce the time decay factor based on a second time step.

4. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 2 are implemented.

5. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 2 are implemented.

Images