Mechanical arm control method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention discloses a mechanical arm control method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring task parameters when a target mechanical arm executes a control task; according to the recurrent neural network, acquiring a joint angle sequence of each joint of the target mechanical arm from an initial position to a target position when the historical task is executed; and according to the task parameters and the historical task joint angle sequence, target joint angles of all joints of the target mechanical arm at the next moment are predicted, and all the joints are controlled to execute task instructions according to the target joint angles. By the adoption of the scheme, the joint angle of the mechanical arm can be flexibly adjusted according to different task requirements by introducing the multi-task parameters, multi-task operation is achieved, and the dynamic environment adaptability is enhanced; and by means of the processing capacity of the recurrent neural network on the time sequence, real-time prediction and adjustment of the joint angle of the mechanical arm are achieved, the operation stability and the task completion rate are improved, and high adaptability and universality are achieved in a multi-task scene.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of robotics technology, and in particular to a robot arm control method, device, electronic device and storage medium. Background Art

[0002] With the development of industrial automation and intelligent manufacturing, robotics technology is widely used in various fields. As the core component of industrial robots, robotic arms are often used in production links such as assembly, welding, handling and spraying. The control system of a robotic arm usually includes joints with multiple degrees of freedom. Through precise control, these joints can complete various complex movements.

[0003] Traditional robotic arm control systems rely on classic control algorithms, such as proportional-integral-differential control and kinematic models. These methods perform well in simple and clear scenarios, but are often insufficient in complex, changing operating environments and multi-task requirements. Existing robotic arm control systems rely on predefined motion trajectories and fixed control algorithms, but these methods have several major problems: poor multi-task adaptability. Traditional methods are optimized for specific tasks and are difficult to adapt to different operating requirements in multi-task environments; insufficient adaptability to dynamic environments. Traditional algorithms cannot be adjusted in real time when facing dynamically changing operating environments, resulting in poor control effects or task failures; complex motion control is difficult, especially for multi-degree-of-freedom robotic arms. Traditional methods have challenges in accurately controlling multiple joint angles and coordinating the movements of each joint, which can easily lead to cumulative errors and control instability; real-time and computational efficiency. Traditional methods are difficult to meet the requirements of real-time and computational efficiency under complex tasks and high-precision control, which limits the application and performance improvement of robotic arms.

[0004] Therefore, how to control a multi-degree-of-freedom robotic arm is a technical problem that needs to be solved urgently by those skilled in the art. Summary of the invention

[0005] The embodiments of the present invention provide a robot arm control method, device, electronic device and storage medium to achieve real-time prediction and adjustment of the robot arm joint angles, improve operational stability and task completion rate, and have strong adaptability and versatility in multi-task scenarios.

[0006] In a first aspect, an embodiment of the present invention provides a method for controlling a robotic arm, comprising:

[0007] Obtaining task parameters when the target manipulator performs a control task; the task parameters include initial joint angles, number of joints, task instructions, connecting rod length, and connecting shaft offset;

[0008] According to the recurrent neural network, the joint angle sequence of each joint of the target robot arm from the initial position to the target position when performing historical tasks is obtained;

[0009] Based on the task parameters and the historical task joint angle sequence, the target joint angle of each joint of the target robot arm at the next moment is predicted, and each joint is controlled to execute the task instruction according to the target joint angle.

[0010] In a second aspect, an embodiment of the present invention further provides a robot arm control device, comprising:

[0011] A task parameter acquisition module is used to obtain the task parameters of the target manipulator when performing the control task; the task parameters include the initial joint angle, the number of joints, the task instruction, the connecting rod length and the connecting shaft offset;

[0012] The joint angle sequence acquisition module is used to obtain the joint angle sequence of each joint of the target robot arm from the initial position to the target position when performing historical tasks based on the recurrent neural network;

[0013] The robot arm angle control module is used to predict the target joint angle of each joint of the target robot arm at the next moment based on the task parameters and the historical task joint angle sequence, and control each joint to execute the task instruction based on the target joint angle.

[0014] In a third aspect, an embodiment of the present invention further provides an electronic device, the electronic device comprising:

[0015] one or more processors;

[0016] A storage device for storing one or more programs;

[0017] When the one or more programs are executed by the one or more processors, the one or more processors implement the robotic arm control method described in any embodiment of the present invention.

[0018] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the robot arm control method described in any embodiment of the present invention.

[0019] In a fifth aspect, an embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the robot arm control method as described in any embodiment of the present invention.

[0020] The embodiment of the present invention provides a robot arm control method, device, electronic device and storage medium, which obtains the initial joint angle, number of joints, task instructions, connecting rod length and connecting shaft offset when the target robot arm performs the control task; obtains the joint angle sequence of each joint of the target robot arm from the initial position to the target position when performing the historical task according to the recurrent neural network; predicts the target joint angle of each joint of the target robot arm at the next moment according to the task parameters and the historical task joint angle sequence, and controls each joint to perform the task instruction according to the target joint angle. The technical solution of the embodiment of the present invention is adopted, and by introducing multi-task parameters, the robot arm can flexibly adjust the joint angle control strategy according to different task requirements, realize multiple task operations, and enhance the adaptability to dynamic environments; the processing ability of the recurrent neural network for time series data is used to realize real-time prediction and adjustment of the joint angle of the robot arm, improve the operational stability and task completion rate, and have strong adaptability and versatility in multi-task scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. Also, the same reference symbols are used throughout the drawings to represent the same parts. In the drawings:

[0022] Figure 1 is a flow chart of a robot arm control method provided in an embodiment of the present invention;

[0023] Figure 2 is a flow chart of another robot arm control method provided in an embodiment of the present invention;

[0024] Figure 3 is a structural schematic diagram of a robotic arm control device provided in an embodiment of the present invention;

[0025] Figure 4 It is a structural schematic diagram of an electronic device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0027] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the operations (or steps) as sequential processes, many of the operations (or steps) therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the accompanying drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0028] Among them, the acquisition, storage, use and processing of data in the technical solution of this application are in compliance with the relevant provisions of national laws and regulations. It should be noted that in the embodiments of this application, some existing solutions in the industry such as certain software, components or models may be mentioned, which should be considered as exemplary, and their purpose is only to illustrate the feasibility of the implementation of the technical solution of this application, but it does not mean that the applicant has or will necessarily use the solution.

[0029] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein.

[0030] Embodiment 1

[0031] Figure 1 1 is a flow chart of a method for controlling a robotic arm provided in an embodiment of the present invention. This embodiment is applicable to the case of controlling a robotic arm. The method of this embodiment can be executed by a robotic arm control device, which can be implemented in hardware and / or software. The device can be configured in a server for controlling a robotic arm. The method specifically includes the following steps:

[0032] S110, obtaining task parameters when the target robotic arm performs a control task.

[0033] Among them, the robotic arm is the core component of industrial robots and is widely used in various industrial operations such as assembly, welding, handling, and spraying. The control system of the robotic arm usually includes joints with multiple degrees of freedom, which can complete various complex movements by precisely controlling these joints.

[0034] The task parameters include initial joint angle, number of joints, task instruction, connecting rod length and connecting shaft offset. The initial joint angle represents the joint angle of each joint of the target robot arm in the initial state, which is expressed as θ0=[θ1, θ2, θ3...θ n ], where θn It represents the initial angle of the nth joint of the robot arm, where n is the degree of freedom of the robot arm. The embodiment of the present invention supports a robot arm with any degree of freedom, such as 6 or 8 degrees of freedom.

[0035] The task instructions are used to describe the specific tasks that the target robot arm needs to perform, including grabbing operation instructions and handling operation instructions. Different tasks correspond to different instruction vectors, and the corresponding vectors can be generated with the help of the semantic vector library in the large model.

[0036] The physical parameters of the robot arm include the connecting rod length and the connecting shaft offset, which are used to describe the specific structure of the robot arm. Among them, the connecting rod length (L) is used to describe the length of each part of the robot arm, expressed as L = [L1, L2, L3...L n ]; where L n Indicates the length of the nth link of the robot. The joint offset (O) is used to describe the offset between the joints, expressed as O = [O1, O2, O3...O n ]; where O n Indicates the offset distance of the nth joint of the robot arm.

[0037] S120. Obtain, based on a recurrent neural network, a joint angle sequence of each joint of the target robotic arm from an initial position to a target position when executing historical tasks.

[0038] Among them, the Recurrent Neural Network (RNN) is a special neural network architecture that is good at processing sequence data and time series prediction. Unlike traditional feedforward neural networks, RNN has a built-in loop structure, which enables it to retain historical information in sequence data. Specifically, the hidden state of RNN can record the information of the previous moment, combine it with the current input, and output the result of the current moment. Therefore, RNN is very suitable for processing tasks that need to consider time dependencies, such as language models, speech recognition, time series prediction, etc. The embodiment of the present invention uses the recurrent neural network to obtain the joint angle sequence of each joint of the target robotic arm from the initial position to the target position when performing historical tasks.

[0039] The joint angle sequence refers to the joint angle sequence corresponding to each joint of the target robotic arm at each time step from the initial position to the target position when performing a task. For example, the target robotic arm includes 6 joints, and the first joint needs 4 time steps to go from the initial position to the target position when performing a task. The sequence composed of the angles of each time step of the first joint from the first time step to the fourth time step is the joint angle sequence of the first joint. The embodiment of the present invention adopts the recurrent neural network to obtain the joint angle sequence of the 6 joints of the target robotic arm, so as to be used in the subsequent prediction of the joint angle.

[0040] S130, predicting the target joint angle of each joint of the target robot arm at the next moment based on the task parameters and the historical task joint angle sequence, and controlling each joint to execute the task instruction based on the target joint angle.

[0041] Among them, the historical task joint angle sequence is obtained through the recurrent neural network, and the joint angles of the target manipulator at different time steps can be analyzed through the joint angle sequence. Therefore, the target joint angles of each joint of the target manipulator at the next moment can be predicted, and the predicted target joint angles are used to control each joint to execute the task instruction. For example, through the joint angle sequence, it can be analyzed that the joint angle of the first joint of the first target manipulator increases by 5° at each time step, and the joint angle of the second joint increases by 4° at each time step. It is known that the joint angle of the first joint at the current moment is 36°, then the target joint angle of the first joint at the next moment is 41°; the joint angle of the second joint at the current moment is 20°, then the target joint angle of the second joint at the next moment is 24°. The first target manipulator is controlled to perform the control operation according to the target joint angles of the first joint and the second joint. Among them, the connecting rod length, the number of joints and the coupling offset of the first target manipulator remain unchanged.

[0042] Optionally, the target joint angles of each joint of the target manipulator at the next moment are different when the task parameters are different. For example, if the connecting rod length of the second target manipulator changes, the joint angle sequence of each joint of the second target manipulator under different connecting rod lengths is determined by a recurrent neural network, so as to infer the target joint angles of each joint at the next moment under different connecting rod lengths.

[0043] The embodiment of the present invention provides a method for controlling a robotic arm, by obtaining the task parameters of the target robotic arm when performing a control task; the task parameters include the initial joint angle, the number of joints, the task instruction, the connecting rod length, and the connecting shaft offset; according to the recurrent neural network, the joint angle sequence of each joint of the target robotic arm from the initial position to the target position when performing a historical task is obtained; according to the task parameters and the historical task joint angle sequence, the target joint angle of each joint of the target robotic arm at the next moment is predicted, and each joint is controlled to perform the task instruction according to the target joint angle. The technical solution of the embodiment of the present invention is adopted, and by introducing multi-task parameters, the robotic arm can flexibly adjust the joint angle control strategy according to different task requirements, realize multiple task operations, and enhance the adaptability to dynamic environments; the processing ability of the recurrent neural network for time series data is utilized to realize real-time prediction and adjustment of the joint angle of the robotic arm, improve the operational stability and task completion rate, and have strong adaptability and versatility in multi-task scenarios.

[0044] Embodiment 2

[0045] Figure 2 Flow chart of a robot arm control method provided in an embodiment of the present invention. The embodiment of the present invention further optimizes the above embodiment on the basis of the above embodiment, and the embodiment of the present invention can be combined with various optional solutions in one or more of the above embodiments. Figure 2 As shown, the robot arm control method provided in the embodiment of the present invention may include the following steps:

[0046] S210, obtaining task parameters when the target robotic arm performs a control task.

[0047] S220. Obtain, based on the recurrent neural network, a joint angle sequence of each joint of the target robotic arm from an initial position to a target position when executing historical tasks.

[0048] S230: Construct a robotic arm angle control model based on a recurrent neural network according to the task parameters and the joint angle sequence.

[0049] Among them, recurrent neural networks have significant advantages in dynamic system modeling and control. For robotic arm control, joint angles and task instructions can be regarded as a sequence of data, in which the joint angles at each time step depend not only on the current task instructions, but also on the joint angles at the previous moment. Through the RNN model, the system can capture and utilize these temporal dependencies, thereby achieving accurate prediction and control of the joint angles of the robotic arm.

[0050] In the embodiment of the present invention, task parameters, task instructions and joint angle sequences are used as input to construct a robotic arm angle control model based on a recurrent neural network to predict the target joint angles of each joint at the next moment.

[0051] Optionally, the model training data includes a joint angle sequence from an initial position to a target position when the robot performs different tasks. The training data can be generated by a simulation environment or by actually operating the robot, and the training data can be expressed as:

[0052]

[0053] in, represents the initial joint angle of the i-th sample, T i Represents the task instruction, P i Represents the physical parameters of the robot.

[0054] The joint angle sequence can be expressed as:

[0055]

[0056] in Represents the joint angles of the robot at different time steps for the i-th sample in time step t.

[0057] In the robotic arm control system, the introduction of RNN provides a new idea for solving the problems of multi-task, multi-degree-of-freedom control and dynamic environment adaptability that are difficult to deal with by traditional control methods. By training the RNN model, the system can gradually adjust the joint angles of the robotic arm according to the time series information after receiving the task instructions, and finally complete the complex task operation. This intelligent control method significantly improves the operation ability and flexibility of the robotic arm, and opens up a new direction for the development of industrial robot technology.

[0058] S240, predicting the target joint angles of each joint of the target robotic arm at the next moment according to the recurrent neural network-based robotic arm angle control model and the task instructions, and controlling each joint to execute the task instructions according to the target joint angles.

[0059] The robot arm angle control model based on the recurrent neural network predicts the target joint angles of each joint of the target robot arm at the next moment when executing the current task instruction based on the historical joint angle information of each joint when executing the historical task instruction saved in the hidden state and the joint angles of each joint at the current moment when executing the current task instruction; wherein the recurrent neural network includes a hidden state, and the hidden state saves the historical joint angle information of each joint of the target robot arm when executing the historical task instruction. The hidden state can be expressed as:

[0060] h t =σ(W h *h t-1 +W x *x t +b h )

[0061] Among them, W h and w x is the weight matrix; b h is the bias; σ is the activation function, such as ReLU or tanh; x t is the joint angle at the current moment; the hidden state h t The current joint angle of the robot and the historical joint angle are combined to ensure that the model captures the necessary time dependencies during task execution. The output of RNN is the joint angle θ at the next moment. t+1 , the calculation formula is:

[0062] θ t+1 =W y *h t +b y

[0063] Among them, Wy is the weight matrix, b y is the bias. The output layer is based on the current hidden state h t Generate the target joint angle of the target robot arm at the next moment.

[0064] This paper proposes a general recurrent neural network architecture suitable for multi-task angle control of a robotic arm. This architecture can adapt to different types of RNNs (such as standard RNNs, LSTMs, etc.), and through the update mechanism of hidden states, it can capture the dynamic changes of the robotic arm in the time series and realize the accurate mapping of task instructions to joint angles.

[0065] As an optional but non-limiting implementation method, the robotic arm angle control model based on the recurrent neural network includes a robotic arm angle control model supporting any degree of freedom, a robotic arm angle control model supporting any connecting rod length, and a robotic arm angle control model supporting any connecting shaft offset.

[0066] The robotic arm angle control model based on a recurrent neural network constructed in an embodiment of the present invention can construct a robotic arm angle control model suitable for multiple scenarios due to different input parameters; including but not limited to a robotic arm angle control model supporting any degree of freedom, a robotic arm angle control model supporting any connecting rod length, and a robotic arm angle control model supporting any connecting shaft offset.

[0067] As an optional but non-limiting implementation, the target joint angles of each joint of the target robotic arm at the next moment are predicted based on the robotic arm angle control model based on the recurrent neural network and the task instructions, and each joint is controlled to execute the task instructions based on the target joint angles, including:

[0068] If the robot arm angle control model based on the recurrent neural network is a robot arm angle control model that supports arbitrary degrees of freedom, then according to the robot arm angle control model and the task instruction, the target joint angles of each joint of the target robot arm with different degrees of freedom at the next moment are predicted, and each joint is controlled to execute the task instruction according to the target joint angles; wherein the number of degrees of freedom represents the number of joints of the target robot arm;

[0069] Or, if the robot angle control model based on the recurrent neural network is a robot angle control model that supports any connecting rod length, then according to the robot angle control model and the task instruction, the target joint angles of each joint of the target robot arm with different connecting rod lengths at the next moment are predicted, and each joint is controlled to execute the task instruction according to the target joint angles;

[0070] Alternatively, if the robot angle control model based on the recurrent neural network is a robot angle control model that supports arbitrary coupling offsets, then based on the robot angle control model and the task instructions, the target joint angles of each joint of the target robot arm with different coupling offsets at the next moment are predicted, and each joint is controlled to execute the task instructions based on the target joint angles.

[0071] Among them, the robot arm angle control model supporting any degree of freedom can refer to predicting the joint angles of each joint of the target robot arm at different degrees of freedom through a recurrent neural network to achieve the expansion of the degrees of freedom, from the traditional 6 degrees of freedom to n degrees of freedom, and support robots with different degrees of freedom, such as 6 degrees of freedom, 8 degrees of freedom, etc. The robot arm angle control model supporting any degree of freedom can output the corresponding number of joint angles at each moment according to the degrees of freedom of the robot arm, ensuring the precise operation of the robot arm in complex tasks.

[0072] The robot arm angle control model supporting any connecting rod length and the robot arm angle control model supporting any coupling offset take into account the physical parameters of the robot arm, including physical characteristics such as connecting rod length and coupling offset. This design enables the model to adapt to robot arms of different specifications and degrees of freedom, and implement a unified control strategy in different task scenarios, ensuring the versatility and flexibility of the system.

[0073] In an optional solution of the embodiment of the present invention, the mean square error (MSE) is used as the loss function to calculate the error between the model predicted angle and the actual angle; the model weights are updated by the back propagation algorithm through time expansion to optimize the robot arm angle control model based on the recurrent neural network. The mean square error loss function can be expressed as:

[0074]

[0075] in, represents the actual joint angle of the i-th sample at time step t, represents the joint angles predicted by the model.

[0076] The time-expanded back-propagation algorithms include Adam and SGD. During the model optimization process, the weight matrix W is gradually adjusted. h , W x and W y , to minimize the loss function.

[0077] In an optional solution of the embodiment of the present invention, after the model training is completed, an independent validation set is used to evaluate the model performance. During the validation, the initial angle, task instructions and physical parameters of the robot arm are input into the model, the joint angle of each time step is predicted, and compared with the actual data to calculate the error.

[0078] The embodiment of the present invention provides a method for controlling a robotic arm. By taking task instructions and joint angle parameters as input, end-to-end control from task instructions to joint angles is achieved, the control process is simplified, and the flexibility and versatility of the system are improved. The powerful time series modeling capability of RNN is utilized to accurately predict the joint angle of the robotic arm at the next moment, realize intelligent dynamic adjustment, and improve control accuracy and stability. The system can respond to environmental changes in real time, predict and adjust by continuously inputting joint angle data, and ensure reliable operation of the robotic arm in a complex dynamic environment. Advanced deep learning training methods, such as back propagation through time and optimization algorithms, are used to efficiently train the RNN model to ensure rapid convergence and optimization of model parameters. The system design has good scalability, and the performance and application scope of the control system are further improved by introducing more complex neural network structures or combining reinforcement learning methods. By optimizing the RNN model structure and adopting efficient computing resource management strategies, the control system is ensured to have good real-time performance and computing efficiency, and to adapt to high-frequency control requirements in industrial production. By introducing the Dropout layer and other regularization techniques, the robustness of the model is improved, the sensitivity to input noise and environmental disturbances is reduced, and the stable operation of the robot arm under complex conditions is ensured; the end-to-end control method reduces the dependence on traditional control algorithms, simplifies the system development and maintenance process, reduces development costs and time, and improves the economic benefits of the system. In summary, the multi-task robot arm angle control system based on recurrent neural networks provided by the present invention overcomes the shortcomings of traditional control methods, significantly improves the operation ability and control accuracy of the robot arm in complex and changeable environments, and has broad application prospects and market value.

[0079] Embodiment 3

[0080] Figure 3 : is a schematic diagram of the structure of a robot arm control device provided in an embodiment of the present invention. The technical solution of this embodiment can be applied to the situation of robot arm control. The device can be implemented by software and / or hardware and is generally integrated on any electronic device with network communication function, including but not limited to: servers, computers, personal digital assistants and other devices. Figure 3 As shown, the robot arm control device provided in this embodiment may include: a task parameter acquisition module 310, a joint angle sequence acquisition module 320 and a robot arm angle control module 330; wherein,

[0081] The task parameter acquisition module 310 is used to acquire the task parameters of the target robot arm when performing the control task; the task parameters include the initial joint angle, the number of joints, the task instruction, the connecting rod length and the connecting shaft offset;

[0082] The joint angle sequence acquisition module 320 is used to acquire the joint angle sequence of each joint of the target manipulator from the initial position to the target position when performing the historical task according to the recurrent neural network;

[0083] The robot arm angle control module 330 is used to predict the target joint angle of each joint of the target robot arm at the next moment based on the task parameters and the historical task joint angle sequence, and control each joint to execute the task instruction based on the target joint angle.

[0084] Based on the above embodiment, optionally, the initial joint angle represents the joint angle of each joint of the target robotic arm in the initial state; the task instruction is used to describe the specific task to be performed by the target robotic arm, including grasping operation instructions and carrying operation instructions.

[0085] Based on the above embodiment, optionally, the robot arm angle control module is specifically used to:

[0086] According to the task parameters and the joint angle sequence, a robotic arm angle control model based on a recurrent neural network is constructed;

[0087] According to the recurrent neural network-based robot arm angle control model and task instructions, the target joint angles of each joint of the target robot arm at the next moment are predicted, and each joint is controlled to execute the task instructions according to the target joint angles.

[0088] On the basis of the above embodiment, optionally, the robot arm angle control module is further specifically used for:

[0089] The robotic arm angle control model based on the recurrent neural network predicts the target joint angles of each joint of the target robotic arm at the next moment when executing the current task instruction based on the historical joint angle information of each joint when executing the historical task instruction saved in the hidden state and the joint angles of each joint at the current moment when executing the current task instruction; wherein, the recurrent neural network includes a hidden state, and the hidden state saves the historical joint angle information of each joint of the target robotic arm when executing the historical task instruction.

[0090] Based on the above embodiments, optionally, the robot arm angle control model based on recurrent neural network includes a robot arm angle control model supporting arbitrary degrees of freedom, a robot arm angle control model supporting arbitrary connecting rod length, and a robot arm angle control model supporting arbitrary connecting shaft offset.

[0091] On the basis of the above embodiment, optionally, the robot arm angle control module is further specifically used for:

[0092] If the robot arm angle control model based on the recurrent neural network is a robot arm angle control model that supports arbitrary degrees of freedom, then according to the robot arm angle control model and the task instruction, the target joint angles of each joint of the target robot arm with different degrees of freedom at the next moment are predicted, and each joint is controlled to execute the task instruction according to the target joint angles; wherein the number of degrees of freedom represents the number of joints of the target robot arm;

[0093] Or, if the robot angle control model based on the recurrent neural network is a robot angle control model that supports any connecting rod length, then according to the robot angle control model and the task instruction, the target joint angles of each joint of the target robot arm with different connecting rod lengths at the next moment are predicted, and each joint is controlled to execute the task instruction according to the target joint angles;

[0094] Alternatively, if the robot angle control model based on the recurrent neural network is a robot angle control model that supports arbitrary coupling offsets, then based on the robot angle control model and the task instructions, the target joint angles of each joint of the target robot arm with different coupling offsets at the next moment are predicted, and each joint is controlled to execute the task instructions based on the target joint angles.

[0095] The robotic arm control device provided in the embodiment of the present invention can execute the robotic arm control method provided in any embodiment of the present invention mentioned above, and has the corresponding functions and beneficial effects of executing the robotic arm control method. For detailed process, please refer to the relevant operations of the robotic arm control method in the above embodiment.

[0096] Embodiment 4

[0097] Figure 4 1 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0098] like Figure 4As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0099] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0100] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a robotic arm control method.

[0101] In some embodiments, the robot arm control method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the robot arm control method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the robot arm control method in any other appropriate manner (e.g., by means of firmware).

[0102] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0103] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0104] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0105] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0106] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0107] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0108] Embodiment 5

[0109] An embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the robot arm control method provided in any embodiment of the present application.

[0110] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0111] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0112] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A robot arm control method, characterized in that: The method comprises: Obtaining task parameters when the target manipulator performs a control task; the task parameters include initial joint angles, number of joints, task instructions, connecting rod length, and connecting shaft offset; According to the recurrent neural network, the joint angle sequence of each joint of the target robot arm from the initial position to the target position when performing historical tasks is obtained; Based on the task parameters and the historical task joint angle sequence, the target joint angle of each joint of the target robot arm at the next moment is predicted, and each joint is controlled to execute the task instruction according to the target joint angle.

2. The method according to claim 1, characterized in that The initial joint angle represents the joint angle of each joint of the target robot arm in the initial state; the task instruction is used to describe the specific task that the target robot arm needs to perform, including grasping operation instructions and carrying operation instructions.

3. The method according to claim 1, characterized in that The method predicts the target joint angle of each joint of the target robot arm at the next moment based on the task parameters and the historical task joint angle sequence, and controls each joint to execute the task instruction based on the target joint angle, including: According to the task parameters and the joint angle sequence, a robotic arm angle control model based on a recurrent neural network is constructed; According to the recurrent neural network-based robot arm angle control model and task instructions, the target joint angles of each joint of the target robot arm at the next moment are predicted, and each joint is controlled to execute the task instructions according to the target joint angles.

4. The method according to claim 3, characterized in that The method predicts the target joint angles of each joint of the target robotic arm at the next moment according to the robotic arm angle control model based on the recurrent neural network and the task instructions, including: The robotic arm angle control model based on the recurrent neural network predicts the target joint angles of each joint of the target robotic arm at the next moment when executing the current task instruction based on the historical joint angle information of each joint when executing the historical task instruction saved in the hidden state and the joint angles of each joint at the current moment when executing the current task instruction; wherein, the recurrent neural network includes a hidden state, and the hidden state saves the historical joint angle information of each joint of the target robotic arm when executing the historical task instruction.

5. The method according to claim 3, characterized in that: The robotic arm angle control model based on the recurrent neural network includes a robotic arm angle control model supporting any degree of freedom, a robotic arm angle control model supporting any connecting rod length, and a robotic arm angle control model supporting any connecting shaft offset.

6. The method according to claim 5, characterized in that The method predicts the target joint angle of each joint of the target robotic arm at the next moment according to the robotic arm angle control model based on the recurrent neural network and the task instruction, and controls each joint to execute the task instruction according to the target joint angle, including: If the robot arm angle control model based on the recurrent neural network is a robot arm angle control model that supports arbitrary degrees of freedom, then according to the robot arm angle control model and the task instruction, the target joint angles of each joint of the target robot arm with different degrees of freedom at the next moment are predicted, and each joint is controlled to execute the task instruction according to the target joint angles; wherein the number of degrees of freedom represents the number of joints of the target robot arm; Or, if the robot angle control model based on the recurrent neural network is a robot angle control model that supports any connecting rod length, then according to the robot angle control model and the task instruction, the target joint angles of each joint of the target robot arm with different connecting rod lengths at the next moment are predicted, and each joint is controlled to execute the task instruction according to the target joint angles; Alternatively, if the robot angle control model based on the recurrent neural network is a robot angle control model that supports arbitrary coupling offsets, then based on the robot angle control model and the task instructions, the target joint angles of each joint of the target robot arm with different coupling offsets at the next moment are predicted, and each joint is controlled to execute the task instructions based on the target joint angles.

7. A robot arm control device, characterized in that: The device comprises: A task parameter acquisition module is used to obtain the task parameters of the target manipulator when performing the control task; the task parameters include the initial joint angle, the number of joints, the task instruction, the connecting rod length and the connecting shaft offset; The joint angle sequence acquisition module is used to obtain the joint angle sequence of each joint of the target robot arm from the initial position to the target position when performing historical tasks based on the recurrent neural network; The robot arm angle control module is used to predict the target joint angle of each joint of the target robot arm at the next moment based on the task parameters and the historical task joint angle sequence, and control each joint to execute the task instruction based on the target joint angle.

8. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the robot arm control method described in any one of claims 1-6.

9. A storage medium containing computer executable instructions, characterized in that: The computer executable instructions are used to execute the robot arm control method as described in any one of claims 1-6 when executed by a computer processor.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the robot arm control method according to any one of claims 1 to 6 is implemented.

Images