Mechanical arm motion trail optimization method and device based on artificial intelligence

Abstract

The invention relates to the technical field of artificial intelligence, and discloses an artificial intelligence-based mechanical arm motion track optimization method, which comprises the following steps of: acquiring motion state data and environment information of a mechanical arm in different environments and motion states through a plurality of groups of sensor units on the mechanical arm; cleaning, standardizing and enhancing the collected comprehensive sensing data; constructing a mixed space training model which comprises a state space and an action space; based on output of the hybrid space training model, an optimal mechanical arm track is generated; and the track of the mechanical arm is dynamically adjusted based on real-time feedback information of the motion state of the mechanical arm. A mixed space training model including a state space and an action space is constructed, the motion state and possible actions of the mechanical arm can be described more comprehensively, and the expression ability and optimization effect of the model are improved.

Description

Technical Field

[0001] The present invention and the field of artificial intelligence technology specifically relate to a method and device for optimizing the motion trajectory of a robotic arm based on artificial intelligence. Background Art

[0002] Robotic arms play a vital role in modern industrial production and service industries. They can perform a variety of complex tasks, such as assembly, handling, welding, and medical surgery. The performance of a robotic arm is directly related to its efficiency and accuracy, and trajectory planning is one of the key factors determining its performance.

[0003] Traditional robotic arm trajectory planning methods are typically based on predefined algorithms, such as polynomial interpolation and spline curves. While these methods can generate smooth trajectories, they lack adaptability and flexibility in complex environments. For example, in environments with obstacles or that require rapid response to changes, traditional methods often struggle to generate optimal trajectories, resulting in reduced robotic arm efficiency or even inability to complete the task. Summary of the Invention

[0004] (1) Technical problems solved

[0005] In response to the deficiencies of the existing technology, the present application provides a method and device for optimizing the motion trajectory of a robotic arm based on artificial intelligence.

[0006] (2) Technical solution

[0007] To solve the above problems, the present application provides the following technical solution: a method for optimizing the motion trajectory of a robotic arm based on artificial intelligence, comprising the following steps:

[0008] The motion state data and environmental information of the robotic arm in different environments and motion states are collected by multiple sensor units on the robotic arm, and the collected sensor data are fused and processed to form comprehensive perception data of the optimization target;

[0009] Cleaning, standardizing, and enhancing the collected comprehensive perception data to convert the comprehensive perception data into a format suitable for model training;

[0010] Build a hybrid space training model, including state space and action space;

[0011] Generate the optimal robot arm trajectory based on the output of the hybrid space training model;

[0012] Based on the real-time feedback information of the robot arm's motion status, the trajectory of the robot arm is dynamically adjusted, and the model parameters are continuously updated through the online learning mechanism.

[0013] Preferably, the plurality of sensor units include angle sensors, velocity sensors, and acceleration sensors, which are respectively installed at the joints of the robotic arm to obtain data of the robotic arm in different motion states;

[0014] An image acquisition device is provided in the working environment of the robotic arm. The image acquisition device may be a high-speed camera, which is used to acquire image information of the robotic arm in different environments. The image information includes the position information of the robotic arm, the movement trajectory information of the robotic arm, the position information of obstacles in the environment, and the shape information of obstacles;

[0015] The status data and environmental information collected by the sensor unit and the image acquisition device are stored in a database.

[0016] Preferably, the objectives to be optimized include at least the robot arm trajectory length, robot arm trajectory smoothness, robot arm obstacle avoidance cost, energy consumption and movement time. A weight is assigned to the optimization objective to reflect the relative importance of the optimization objective in the overall optimization process, and the optimization objectives are weighted and summed. The calculation formula is as follows:

[0017] J=W1*L+W2*S+W3*C+W4*E+W5*T

[0018] Among them, L is the trajectory length, S is the trajectory smoothness, C is the obstacle avoidance cost, E is the energy consumption, T is the movement time, W1, W2, W3, W4, and W5 are the weights of the corresponding targets, which are used to adjust the relative importance of different targets.

[0019] Preferably, the data cleaning includes filters or statistical methods to remove noise from the data;

[0020] The data standardization is to standardize different types of data so that they have the same scale;

[0021] The data enhancement increases the amount of data by adding noise, rotating or scaling, thereby improving the generalization ability of the model.

[0022] Preferably, the state space is used to describe the complete state of the manipulator at any given time. The state space includes the position, velocity and acceleration of the joints of the manipulator. The state vector s can be expressed as:

[0023] S=[θ1,θ2,…,θ n ,θ1˙,θ2˙,…,θn˙,θ1¨,θ2¨,…,θ n ¨]

[0024] Among them, θ i is the angle of the i-th joint, θ i˙ is the angular velocity of the i-th joint, θ i ¨ is the angular acceleration of the i-th joint;

[0025] The action space is used to describe all possible actions that the robot can perform in a given state. The action vector a is expressed as:

[0026] a=[Δθ1,Δθ2,…,Δθn,Δθ1˙,Δθ2˙,…,Δθ n ˙,Δθ1¨,

[0027] Δθ2¨,…,Δθ n ¨]

[0028] Where Δθ i is the angle change of the i-th joint, Δθ i ˙ is the velocity change of the i-th joint, Δθ i ¨ is the acceleration change of the i-th joint.

[0029] Preferably, the hybrid space training model is optimized and calculated using a particle swarm optimization method, wherein the particle optimization speed calculation formula is:

[0030] vi(t+1)=w*vi(t)+c1*r1*(pi-xi(t))+c2*r2*(pg-xi(t))

[0031] where vi(t) is the velocity of particle i at time t, w is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers, pi is the individual optimal position of particle i, and pg is the global optimal position of the group.

[0032] Preferably, the trajectory optimization module includes trajectory generation, simulation verification and actual testing;

[0033] The trajectory generation generates the optimal robot arm trajectory by mixing the output of the space training model;

[0034] The simulation verification verifies whether the generated trajectory meets the action requirements through simulation software;

[0035] The actual test verifies the effect of the generated trajectory by testing it in a real environment.

[0036] Preferably, the real-time adjustment module includes two submodules: feedback control and online learning;

[0037] The feedback control dynamically adjusts the trajectory of the robot arm according to real-time feedback information during the execution of the task, ensuring that the robot arm can maintain efficient and precise operation in a changing environment;

[0038] The online learning updates the model parameters through the online learning mechanism, wherein the calculation formula of the online learning is:

[0039]

[0040] Where η is the learning rate.

[0041] Preferably, it further includes a control module, which converts the optimized trajectory into a control instruction for the robotic arm and drives the robotic arm to perform the task. The control module includes trajectory conversion, control signal sending and execution monitoring;

[0042] The trajectory conversion is used to convert the optimized trajectory into control instructions for the robotic arm;

[0043] The control signal is sent by sending a control instruction to the controller of the robotic arm to drive the robotic arm to perform the task;

[0044] The execution monitoring is used to monitor the execution status of the robotic arm.

[0045] Preferably, an artificial intelligence-based robotic arm comprises: at least one robotic arm body and at least one data processor.

[0046] Preferably, a robot arm motion trajectory optimization device based on artificial intelligence is characterized by comprising:

[0047] A data acquisition module, for collecting motion state data and environmental information of the robotic arm in different environments and motion states;

[0048] Data preprocessing module, used to convert collected data into a format suitable for model training;

[0049] The spatial training module trains the model in state space and action space to find the optimal trajectory;

[0050] The trajectory optimization module generates the optimal robot arm trajectory through the output of the spatial training model and verifies it in simulation software and actual environment;

[0051] The real-time adjustment module dynamically adjusts the trajectory of the robotic arm based on the real-time feedback information of the robotic arm's motion status.

[0052] (3) Beneficial effects

[0053] Compared with the existing technology, this application provides a method and device for optimizing the motion trajectory of a robotic arm based on artificial intelligence, which has the following beneficial effects:

[0054] 1. This artificial intelligence-based method and device for optimizing the motion trajectory of a robotic arm constructs a hybrid space training model consisting of a state space and an action space, which can more comprehensively describe the motion state and possible actions of the robotic arm, thereby improving the model's expressive power and optimization effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a flow chart of the artificial intelligence-based robot arm trajectory optimization method for this application;

[0056] Figure 2 This is the architecture diagram of the artificial intelligence-based robotic arm trajectory optimization system for this application;

[0057] Figure 3 This is a schematic diagram of the data collection module of this application;

[0058] Figure 4 This is a structural diagram of the data preprocessing module of this application;

[0059] Figure 5 This is a schematic diagram of the structure of the hybrid space training module of this application;

[0060] Figure 6 This is a schematic diagram of the structure of the trajectory optimization module of this application;

[0061] Figure 7 This is a schematic diagram of the structure of the real-time adjustment module for this application;

[0062] Figure 8 This is a structural diagram of the control module of this application. DETAILED DESCRIPTION

[0063] The present application is further described in detail below by means of specific embodiments in conjunction with the accompanying drawings. Similar elements in different embodiments are numbered with associated similar elements. In the following embodiments, many detailed descriptions are provided to enable the present application to be better understood. However, those skilled in the art will readily appreciate that some of the features may be omitted in different circumstances, or may be replaced by other elements, materials, or methods. In some cases, some operations related to the present application are not shown or described in the specification. This is to avoid the core portion of the present application being overwhelmed by excessive descriptions, and for those skilled in the art, it is not necessary to describe these related operations in detail. They can fully understand the related operations based on the description in the specification and the general technical knowledge in the art.

[0064] In addition, the features, operations, or characteristics described in the specification may be combined in any appropriate manner to form various embodiments. Furthermore, the steps or actions in the method description may be reordered or adjusted in a manner readily apparent to those skilled in the art. Therefore, the various sequences in the specification and drawings are provided solely for the purpose of clearly describing a particular embodiment and are not intended to be mandatory, unless otherwise specified.

[0065] The serial numbers assigned to components herein, such as "first," "second," etc., are used solely to distinguish the objects being described and do not convey any sequential or technical meaning. References to "connection" and "coupling" herein, unless otherwise specified, include both direct and indirect connections (couplings).

[0066] See also Figure 1 and Figure 3 , this application provides a new technical solution: a method for optimizing the motion trajectory of a robotic arm based on artificial intelligence, comprising the following steps:

[0067] Data collection: The motion state data and environmental information of the robotic arm in different environments and motion states are collected through multiple sets of sensor units on the robotic arm, and the collected sensor data are fused and processed to form comprehensive perception data of the optimization target.

[0068] The plurality of sensor units include angle sensors, velocity sensors, and acceleration sensors, which are respectively installed at the joints of the robotic arm to obtain data of the robotic arm in different motion states;

[0069] An image acquisition device is provided in the working environment of the robotic arm. The image acquisition device may be a high-speed camera, which is used to acquire image information of the robotic arm in different environments. The image information includes the position information of the robotic arm, the movement trajectory information of the robotic arm, the position information of obstacles in the environment, and the shape information of obstacles;

[0070] The state data and environmental information collected by the sensor unit and image acquisition device are stored in a database. The optimization targets include at least the robot arm trajectory length, robot arm trajectory smoothness, robot arm obstacle avoidance cost, energy consumption, and movement time. A weight is assigned to each optimization target to reflect its relative importance in the overall optimization process. The optimization targets are then weighted and summed. The calculation formula is as follows:

[0071] J=W1*L+W2*S+W3*C+W4*E+W5*T

[0072] Among them, L is the trajectory length, S is the trajectory smoothness, C is the obstacle avoidance cost, E is the energy consumption, T is the motion time, W1, W2, W3, W4, W5 are the weights of the corresponding targets, which are used to adjust the relative importance of different targets. Furthermore, the calculation of the obstacle avoidance cost C: the obstacle avoidance cost can be evaluated by calculating the minimum distance between the robot arm and the obstacle, or the potential field method can be used to calculate the obstacle avoidance cost; the calculation of the energy consumption E: the energy consumption can be evaluated by calculating the energy consumption of each joint, which is usually related to the moment of inertia and angular velocity of the joint; the calculation of the motion time T: the motion time can be evaluated by calculating the execution time of each point on the trajectory, which is usually related to the maximum speed and acceleration limit of the robot arm; the selection of the weight Wi: the selection of the weight depends on the relative importance of each target in the actual application, which can be determined by experiments or expert experience.

[0073] See also Figure 4 ,In some embodiments, the data pre-processing module includes data cleaning, data ,cleaning, and data enhancement;

[0074] The data cleaning includes filters (such as Kalman filters) or statistical methods to remove noise in the data; for missing data points, interpolation methods or the value of the previous valid data point can be used to fill in the missing data points.

[0075] Data cleaning involves normalizing different types of data to make them have the same scale; for example, converting the data into a distribution with a mean of 0 and a standard deviation of 1. This can be achieved using the following formula:

[0076] z=σx-μ

[0077] Where x is the original data, z is the standardized data, μ is the mean, and σ is the standard deviation;

[0078] The data enhancement increases the amount of data by adding noise, rotating or scaling, thereby improving the generalization ability of the model.

[0079] See also Figure 5 ,In some embodiments, the hybrid space training module includes a state space and an action space;

[0080] The state space is used to describe the complete state of the manipulator at any given time. The state space includes the position, velocity and acceleration of the joints of the manipulator. The state vector s can be expressed as:

[0081] S=[θ1,θ2,…,θ n ,θ1˙,θ2˙,…,θn˙,θ1¨,θ2¨,…,θ n ¨]

[0082] Among them, θ i is the angle of the i-th joint, θi ˙ is the angular velocity of the i-th joint, θ i ¨ is the angular acceleration of the i-th joint;

[0083] The action space is used to describe all possible actions that the robot can perform in a given state. The action vector a is expressed as:

[0084] a=[Δθ1,Δθ2,…,Δθn,Δθ1˙,Δθ2˙,…,Δθ n ˙,Δθ1¨,

[0085] Δθ2¨,…,Δθ n ¨]

[0086] Where Δθ i is the angle change of the i-th joint, Δθ i ˙ is the velocity change of the i-th joint, Δθ i ¨ is the acceleration change of the i-th joint.

[0087] Furthermore, the hybrid space training model is optimized by particle swarm optimization, where the particle optimization speed calculation formula is:

[0088] vi(t+1)=w*vi(t)+c1*r1*(pi-xi(t))+c2*r2*(pg-xi(t))

[0089] where vi(t) is the velocity of particle i at time t, w is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers, pi is the individual optimal position of particle i, and pg is the global optimal position of the group.

[0090] See also Figure 6 ,In some embodiments, the trajectory optimization module includes trajectory generation, simulation verification, and actual testing;

[0091] The trajectory generation generates the optimal robot arm trajectory by mixing the output of the space training model;

[0092] The simulation verification verifies whether the generated trajectory meets the action requirements through simulation software; simulation software includes MATLAB / Simulink, Gazebo, etc.

[0093] The actual test verifies the effectiveness of the generated trajectory by testing it in a real environment. During the test, attention should be paid to the following aspects: safety: ensuring that the robot arm does not cause damage to equipment or personnel during the test; stability: evaluating the stability of the generated trajectory in different environments; and recording the time and energy required for the robot arm to complete the task to evaluate its efficiency.

[0094] See also Figure 7,In some embodiments, the real-time adjustment module includes two sub-modules of feedback ,control and online learning;

[0095] The feedback control dynamically adjusts the trajectory of the robot arm based on real-time feedback information (such as sensor data, visual information, etc.) during the robot arm's task execution, ensuring that the robot arm can maintain efficient and precise operation in a changing environment;

[0096] The online learning updates the model parameters through the online learning mechanism, wherein the calculation formula of the online learning is:

[0097]

[0098] Wherein, η is the learning rate, and the model parameters include weight parameters, particle swarm optimization parameters, state space parameters, action space parameters, and other various values ​​and settings used in artificial intelligence algorithms to describe and optimize the motion trajectory of the robotic arm.

[0099] See also Figure 8 , in some embodiments, further comprising a control module, wherein the control module converts the optimized trajectory into a control instruction of the robotic arm and drives the robotic arm to perform a task, and the control module includes trajectory conversion, control signal sending, and execution monitoring;

[0100] The trajectory conversion is used to convert the optimized trajectory into control instructions for the robotic arm;

[0101] The control signal is sent by sending a control instruction to the controller of the robotic arm to drive the robotic arm to perform the task;

[0102] The execution monitoring is used to monitor the execution status of the robotic arm.

[0103] In some embodiments, an artificial intelligence-based robotic arm includes: at least one robotic arm body and at least one data processor.

[0104] In some embodiments, a robot arm motion trajectory optimization device based on artificial intelligence is characterized by comprising:

[0105] A data acquisition module, for collecting motion state data and environmental information of the robotic arm in different environments and motion states;

[0106] Data preprocessing module, used to convert collected data into a format suitable for model training;

[0107] The spatial training module trains the model in state space and action space to find the optimal trajectory;

[0108] The trajectory optimization module generates the optimal robot arm trajectory through the output of the spatial training model and verifies it in simulation software and actual environment;

[0109] The real-time adjustment module dynamically adjusts the trajectory of the robotic arm based on the real-time feedback information of the robotic arm's motion status.

[0110] Those skilled in the art will appreciate that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer program. When all or part of the functions in the above embodiments are implemented by computer program, the program can be stored in a computer-readable storage medium, and the storage medium can include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to implement the above functions. For example, the program is stored in the memory of the device, and when the program in the memory is executed by the processor, all or part of the above functions can be implemented. In addition, when all or part of the functions in the above embodiments are implemented by computer program, the program can also be stored in a storage medium such as a server, another computer, disk, optical disk, flash disk or mobile hard disk, and saved in the memory of the local device by downloading or copying, or the system of the local device is updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be implemented.

[0111] The above specific examples are used to illustrate the present application, which is only used to help understand the present application and is not intended to limit the present application. For those skilled in the art of the present application, based on the concept of the present application, they can also make some simple deductions, modifications or substitutions.

Claims

1. A method for optimizing the motion trajectory of a robotic arm based on artificial intelligence, characterized in that: The following steps are involved: The motion state data and environmental information of the robotic arm in different environments and motion states are collected by multiple sensor units on the robotic arm, and the collected sensor data are fused and processed to form comprehensive perception data of the optimization target; Cleaning, standardizing, and enhancing the collected comprehensive perception data to convert the comprehensive perception data into a format suitable for model training; Build a hybrid space training model, including state space and action space; Generate the optimal robot arm trajectory based on the output of the hybrid space training model; Based on the real-time feedback information of the robot arm's motion status, the trajectory of the robot arm is dynamically adjusted, and the model parameters are continuously updated through the online learning mechanism.

2. The method for optimizing the motion trajectory of a robotic arm based on artificial intelligence according to claim 1, characterized in that: The plurality of sensor units include angle sensors, velocity sensors, and acceleration sensors, which are respectively installed at the joints of the robotic arm to obtain data of the robotic arm in different motion states; An image acquisition device is provided in the working environment of the robotic arm, wherein the image acquisition device is a high-speed camera, which is used to acquire image information of the robotic arm in different environments, wherein the image information includes the position information of the robotic arm, the motion trajectory information of the robotic arm, the position information of obstacles in the environment, and the shape information of obstacles; The status data and environmental information collected by the sensor unit and the image acquisition device are stored in a database.

3. The method for optimizing the motion trajectory of a robotic arm based on artificial intelligence according to claim 1, characterized in that: The objectives to be optimized include at least the robot arm trajectory length, robot arm trajectory smoothness, robot arm obstacle avoidance cost, energy consumption and movement time. A weight is assigned to the optimization objective to reflect the relative importance of the optimization objective in the overall optimization process, and the optimization objectives are weighted and summed. The calculation formula is as follows: J=W1*L+W2*S+W3*C+W4*E+W5*T Among them, L is the trajectory length, S is the trajectory smoothness, C is the obstacle avoidance cost, E is the energy consumption, T is the movement time, W1, W2, W3, W4, and W5 are the weights of the corresponding targets, which are used to adjust the relative importance of different targets.

4. The method for optimizing the motion trajectory of a robotic arm based on artificial intelligence according to claim 1, wherein: The hybrid space training module includes a state space and an action space; The state space is used to describe the complete state of the manipulator at any given time. The state space includes the position, velocity and acceleration of the joints of the manipulator. The state vector S can be expressed as: S=[θ1, θ2, …, θn, θ1˙, θ2˙, …, θn˙, θ1¨, θ2¨, …, θn¨] Where θi is the angle of the i-th joint, θi˙ is the angular velocity of the i-th joint, and θi¨ is the angular acceleration of the i-th joint; The action space is used to describe all possible actions that the robot can perform in a given state. The action vector a is expressed as: a=[Δθ1,Δθ2,…,Δθn,Δθ1˙,Δθ2˙,…,Δθn˙,Δθ1¨,Δθ2¨,…,Δθn¨] Where Δθi is the angle change of the i-th joint, Δθi˙ is the velocity change of the i-th joint, and Δθi¨ is the acceleration change of the i-th joint.

5. The method for optimizing the motion trajectory of a robotic arm based on artificial intelligence according to claim 4, characterized in that: The hybrid space training model adopts particle swarm optimization to optimize the calculation, where the particle optimization speed calculation formula is: vi(t+1)=w*vi(t)+c1*r1*(pi-xi(t))+c2*r2*(pg-xi(t)) where vi(t) is the velocity of particle i at time t, w is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers, pi is the individual optimal position of particle i, and pg is the global optimal position of the group.

6. The method for optimizing the motion trajectory of a robotic arm based on artificial intelligence according to claim 1, characterized in that: The trajectory optimization module includes trajectory generation, simulation verification and actual testing; The trajectory generation generates the optimal robot arm trajectory by mixing the output of the space training model; The simulation verification verifies whether the generated trajectory meets the action requirements through simulation software; The actual test verifies the effect of the generated trajectory by testing it in a real environment.

7. The method for optimizing the motion trajectory of a robotic arm based on artificial intelligence according to claim 1, characterized in that: The real-time adjustment module includes two submodules: feedback control and online learning; The feedback control dynamically adjusts the trajectory of the robot arm according to real-time feedback information during the execution of the task, ensuring that the robot arm can maintain efficient and precise operation in a changing environment; The online learning updates the model parameters through the online learning mechanism, wherein the calculation formula of the online learning is: Where η is the learning rate.

8. The method for optimizing the motion trajectory of a robotic arm based on artificial intelligence according to claim 1, characterized in that: It also includes a control module, which converts the optimized trajectory into a control instruction for the robotic arm and drives the robotic arm to perform the task. The control module includes trajectory conversion, control signal sending and execution monitoring; The trajectory conversion is used to convert the optimized trajectory into control instructions for the robotic arm; The control signal is sent by sending a control instruction to the controller of the robotic arm to drive the robotic arm to perform the task; The execution monitoring is used to monitor the execution status of the robotic arm.

9. A robotic arm based on artificial intelligence, characterized in that: include: At least one robotic arm body and at least one data processor.

10. A robot arm motion trajectory optimization device based on artificial intelligence, characterized in that: include: A data acquisition module, for collecting motion state data and environmental information of the robotic arm in different environments and motion states; Data preprocessing module, used to convert collected data into a format suitable for model training; The spatial training module trains the model in state space and action space to find the optimal trajectory; The trajectory optimization module generates the optimal robot arm trajectory through the output of the spatial training model and verifies it in simulation software and actual environment; The real-time adjustment module dynamically adjusts the trajectory of the robotic arm based on the real-time feedback information of the robotic arm's motion status.

Images