Method and system for remote control of robotic arms based on artificial intelligence and robotics technology

Abstract

The present invention discloses a method and system for remote control of a robotic arm based on artificial intelligence and robotics technology, which relates to the field of automated control technology, including building a motion model of the robotic arm by integrating an intelligent car lower computer, a high-torque servo motor and a fuzzy control algorithm, and using a high-fidelity 3D pre-planning module for path optimization. The method includes natural language processing command conversion, reinforcement learning optimization path planning, and remote control execution. The entire interactive process realizes closed-loop control from user command input to precise movement of the robotic arm. The method improves the intelligence, flexibility and accuracy of the operation, reduces the complexity of the operation, improves the work efficiency and safety, and provides an efficient and intuitive solution for remote control of the robotic arm.

Description

Technical Field

[0001] The present invention relates to the field of automation control technology, and in particular to a method and system for remotely controlling a robotic arm based on artificial intelligence and robotics technology. Background Art

[0002] In today's era of rapid technological development, robotic arm remote control technology, as an important part of robotics technology, has demonstrated its unique application value in many fields such as industrial automation, medical surgery, and space exploration. In recent years, with the continuous advancement of artificial intelligence technology, especially the development of machine learning, deep learning and other algorithms, the control accuracy and intelligence level of robotic arms have been significantly improved. In the existing technology, the remote control of robotic arms mainly relies on preset programs or manual remote control, which to a certain extent limits the operational flexibility and environmental adaptability of robotic arms.

[0003] However, the existing related technologies still have certain shortcomings. First, the control method of the traditional robot arm often requires a lot of manual intervention when handling complex tasks, and the operation efficiency is low. Secondly, most of the existing robot arm motion models fail to fully consider environmental factors and motion modal parameters, resulting in the stability and accuracy of the robot arm being affected in practical applications. In addition, the path planning method of the prior art often finds it difficult to find the optimal operation path when encountering complex environments, and the real-time and accuracy of command transmission during remote control need to be improved. In response to these problems, the robot arm remote control method based on artificial intelligence and robotics technology invented by us, by introducing an intelligent car lower computer, a high-torque servo motor, a combination of fuzzy control algorithm and adaptive control, and a reinforcement learning algorithm to optimize path planning, is expected to bring significant improvements in operational flexibility, environmental adaptability, and control accuracy. On this basis, our invention also innovatively proposes a semantic understanding model combined with natural language processing, making the remote control of the robot arm more convenient and intelligent. Summary of the invention

[0004] In view of the above existing problems, the present invention aims to solve the problems of remote control accuracy, flexibility and intelligence of the robot arm in a complex environment. By integrating advanced technologies such as high-torque servo motors, fuzzy control algorithms, adaptive control, cameras, power sensors, and using artificial intelligence methods such as natural language processing, high-fidelity 3D preview, RRT path planning algorithm and reinforcement learning, the following goals are achieved: to achieve precise control, efficient path planning and environmental adaptability of the robot arm, improve the convenience and intelligence level of remote control, ensure the safety and reliability of the robot arm when performing tasks, and thus meet the high requirements for remote control of the robot arm in various fields.

[0005] In order to solve the above technical problems, a remote control method of a robotic arm based on artificial intelligence and robotics technology is proposed, including:

[0006] The robot arm is built in the developed intelligent car lower computer, and the robot arm is remotely controlled through the car behavior control module; the improved robot arm is built through the servo motor and control board, and the robot arm motion parameters are collected, the robot arm joint and the end effector position relationship are connected, and the robot arm motion model is built, including the joint angle change, the dynamic relationship between the joint speed and acceleration, and the end effector position; the robot arm motion model is input into the high-fidelity 3D pre-planning module for preview and optimization, and the RRT algorithm is used to generate potential motion paths. By observing the motion trajectory of the robot arm, the path planning is adjusted; the command transmission mechanism between the intelligent car upper computer and the lower computer is designed as In the master-slave mode, the upper computer acts as the master and the lower computer acts as the slave. Operation instructions are input through the upper computer, and a semantic understanding model based on natural language processing is established to convert the user's natural language instructions into robot arm operation instructions, which are transmitted to the lower computer through the network. The BERT machine learning model is applied for semantic understanding. The lower computer parses the command and converts it into motion control instructions, including the rotation angle and movement distance of each axis of the robot arm. The reinforcement learning algorithm is used to optimize and adjust the path planning of the robot arm by adjusting the reward function to find the optimal operation path. The preview and optimization results of the high-fidelity 3D pre-planning module are input into the robot arm control module for remote control.

[0007] As a preferred solution of the method for remote control of a robotic arm based on artificial intelligence and robotics technology described in the present invention, wherein: the construction of the improved robotic arm includes using a high-torque servo motor as a driving unit of the robotic arm, and dynamically adjusting the output of the motor by combining a fuzzy control algorithm with adaptive control;

[0008] The front end of the robot arm is equipped with a camera and a force sensor to detect objects and grasping force, and communicate with the robot's lower computer through an efficient single-chip microcomputer and an embedded control board;

[0009] The motion parameters of the robot arm include motion state parameters, environmental parameters and motion modal parameters;

[0010] The motion state parameters include joint rotation angle, joint rotation speed, and joint rotation acceleration;

[0011] The environmental parameters include the weight of the object carried by the end effector, the effect of temperature changes in the working environment on the motion performance, and the magnitude of friction between the joints;

[0012] The motion mode parameters include movement mode, control mode including autonomous control and remote control, and the initial state of the robot arm before starting to move.

[0013] As a preferred solution of the method for remote control of a robot arm based on artificial intelligence and robotics technology described in the present invention, wherein: the construction of the robot arm motion model includes defining the relationship between the motion reference system of the robot arm and the position of the end effector, and the change amount of the joint angle in a specific time period is fed back by the feedback control amount and the environmental adaptation amount.

[0014]

[0015] in, is the angle of the j-1 joint, the previous joint of the j-th joint, at the previous moment, is the weight coefficient of the control error, is the control error of the jth joint, calculated as the difference between the desired angle and the current angle; is the weight coefficient of environmental adaptation factors, is the environmental parameter influence, caused by load or friction factors; j is the variable index;

[0016] According to the change of joint angle, the dynamic relationship between joint velocity and acceleration is defined:

[0017]

[0018] in, is the velocity of the jth joint at time t, is the time interval, indicating the time period during which the joint displacement occurs; is the velocity of the previous joint j-1, is the influence factor of the previous state on the current speed, controlling the dynamic relationship between the previous and next states;

[0019] Define the end effector position as:

[0020]

[0021] Where P(t) is the position vector of the end effector at time t, The starting position of the robot arm, The length of the jth joint, The adjustment coefficient is used to amplify the impact of environmental factors on the end position; n is the number of joints in the robot arm.

[0022] As a preferred solution of the method for remote control of a robotic arm based on artificial intelligence and robotics technology of the present invention, wherein: the high-fidelity 3D pre-planning module performs preview and optimization, including inputting the motion model of the robotic arm into the high-fidelity 3D pre-planning module, integrating the calculated parameters in the motion model of the robotic arm to form a complete state space, at which point all data are consistent with the real-time feedback of the servo motor and the control board;

[0023] The RRT algorithm is used to generate potential motion paths to ensure that the paths will not collide with obstacles in the virtual environment. The generated paths are executed in the simulation environment. By observing the motion trajectory of the robot arm, the repeatability deviation of the path of the robot arm movement is kept less than 2% in five consecutive simulations. The feasibility and efficiency of the path are evaluated, and the model parameters are adjusted for reverse optimization. When the efficiency of the previewed motion trajectory after optimization is improved by less than 15%, the deep learning algorithm is automatically enabled to iteratively adjust the path planning, and the number of iterations for each adjustment does not exceed 3 times; when the deviation between the actual path and the ideal path exceeds 5 cm, the system will trigger an alarm and require manual adjustment;

[0024] When the user issues a new operation instruction, the host computer should automatically parse and update the motion model through the natural language processing module, re-rehearse, save each rehearsal record, and output the final optimized result to the robot arm control module.

[0025] As a preferred solution of the method for remote control of a robotic arm based on artificial intelligence and robotics technology described in the present invention, the semantic understanding model based on natural language processing includes: the user issues a natural language instruction through voice recognition or text input of an input device, removes stop words through preliminary cleaning, uses an improved word segmentation algorithm to identify semantic units, and constructs a dictionary containing professional terms, verbs, nouns and common phrases for the robotic arm;

[0026] Extract terms from the cleaned text and match them with words in the dictionary. Use a greedy algorithm to match the longest term word by word from left to right. After matching, use natural language processing tools to tag the extracted words with parts of speech, further identify nouns, verbs, and adjectives, and use an improved word segmentation algorithm to build the final operation instructions: Apply BERT's machine learning model for semantic understanding, extract intent and parameters from the instructions, and define new functions ,in, is the natural language representing the input, O is the parsed operation instruction form, and the extracted intent and parameters are mapped to the operation instruction format of the robot arm by defining the transformation matrix T: , where, C, is the robot arm motion command;

[0027] The TCP / IP protocol is used to send the constructed operation instructions to the lower computer through the network. The lower computer parses the command and converts it into motion control instructions. The fixed proportional relationship obtained based on the motion model of the robot arm ,in, Represents the dynamic proportional factor between joint i and joint j, quantifying the influence of joint j on the movement of joint i when it rotates to; Represents the static proportional factor between joint i and joint j, that is, the initial value without dynamic adjustment; Indicates the current angle of joint j How does it affect the scale factor when reaching or exceeding 30 degrees? Function of

[0028]

[0029] in, is the current angle of joint j, An adjustment factor to control the degree of nonlinearity of the influence.

[0030] As a preferred solution of the method for remote control of a robotic arm based on artificial intelligence and robotics technology described in the present invention, the semantic understanding model based on natural language processing also includes that when joint i needs to be adjusted, it is distinguished into joints x related to joint i and joints y not related to joint i. When it is a related joint x, the moving distance of the related joint x is adjusted to:

[0031]

[0032] in, represents the moving distance of the i-th joint, represents the basic moving distance of the i-th joint, that is, the initial distance of the current joint without any additional adjustment; Used to indicate whether the angle of joint j reaches 30 degrees. When it exceeds 30 degrees, When it is 1, the adjustment logic is activated; when it does not reach 30 degrees, 0, no adjustment;

[0033] When it is the related joint y, the angle remains the value given by the instruction:

[0034]

[0035] in, Represents the angle change of joint j; the adjusted moving distance and angle are combined to generate motion instructions:

[0036]

[0037] in, is the joint rotation angle, The distance the joint moves, represents the initial distance before adjustment.

[0038] As a preferred solution of the method for remote control of a robotic arm based on artificial intelligence and robotics technology described in the present invention, the optimal operation path includes real-time feedback of the current state when the user inputs a command, and displaying encouraging information. Based on the feedback, the user is allowed to dynamically adjust the parameters b and c of the reward function to select the optimal operation path:

[0039]

[0040] in, It represents the path with the largest cumulative reward among all possible paths; S represents all possible paths, which is the entire sequence of actions performed by the robot; represents the accumulation from time t=0 to t=T, indicating that the reward is calculated for each time step in the entire path; Indicates the current state The higher the penalty, the farther the current state is from the target. Indicates the current state The smoothness reward, the higher the reward, the smoother the path;

[0041] The preview and optimization results of the high-fidelity 3D pre-planning module are input into the robotic arm control module for remote control, and the instructions are executed according to the optimal operation path.

[0042] Another object of the present invention is to provide a remote control system for a robotic arm based on artificial intelligence and robotics technology. The present invention aims to achieve efficient and accurate remote control of a robotic arm through artificial intelligence and robotics technology. Users can operate the robotic arm through natural language instructions, and the system will optimize its motion path in real time to improve operational efficiency and safety.

[0043] As a preferred solution of the robot arm remote control system based on artificial intelligence and robotics technology described in the present invention, it is characterized by comprising a robot arm construction module, a robot arm preview module, and a robot arm control module;

[0044] The robot arm building module includes a data acquisition module and a motion model building module. The data acquisition module regularly transmits the collected parameters to the motion model building module. The motion model building module analyzes the parameters, builds a motion model of the robot arm, defines the dynamic relationship between the joint angle change, velocity and acceleration, and feeds the motion model back to the data acquisition module for calibration;

[0045] The robot arm preview module includes a high-fidelity 3D pre-planning module, an instruction conversion module and a path optimization module. When the user issues a work instruction, the instruction conversion module converts the natural language instruction into a robot arm operation instruction, parses it into a motion control instruction, inputs it into the high-fidelity 3D pre-planning module, uses the motion model to generate and optimize the motion path of the robot arm, uses the RRT algorithm to ensure the path is safe and effective, and then transmits the optimization result back to the path optimization module for deep learning adjustment to improve the working efficiency of the robot arm;

[0046] The robot control module receives the final optimized path and control instructions from the robot preview module, controls the movement of the robot in real time, and returns the current state to the robot preview module through a feedback mechanism to further optimize the path.

[0047] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and is characterized in that when the processor executes the computer program, the steps of the method for remotely controlling a robotic arm based on artificial intelligence and robotics technology are implemented.

[0048] A computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps of the method for remotely controlling a robotic arm based on artificial intelligence and robotics technology are implemented.

[0049] Beneficial effects of the invention: The invention integrates the robot arm into the lower computer of the intelligent trolley and uses the trolley behavior control module for remote control, thereby realizing the flexible movement and wide operating range of the robot arm. At the same time, the motion model constructed by the high-torque servo motor and fuzzy control algorithm improves the motion accuracy of the robot arm and its adaptability to complex environments, significantly improving the convenience, applicability and stability of the operation of the robot arm.

[0050] In the high-fidelity 3D pre-planning module, we preview and optimize the robot's motion model, use the RRT algorithm to generate potential collision-free motion paths, and combine natural language processing technology to efficiently convert user instructions into robot arm operation instructions; this not only ensures the efficiency and feasibility of path planning, but also simplifies the interaction between the user and the robot, making the operation more intuitive and convenient.

[0051] The reward function is dynamically adjusted through the reinforcement learning algorithm to optimize the path planning of the robot arm, find the optimal operation path, and apply the preview and optimization results to the robot arm control module to achieve precise remote control. This not only improves the operating efficiency and safety of the robot arm, but also enhances the overall coordination of the system and the real-time control, bringing users a higher operating experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0053] Figure 1 An overall flow chart of a method for remotely controlling a robotic arm based on artificial intelligence and robotics technology provided for one embodiment of the present invention.

[0054] Figure 2 A system solution module diagram of a robotic arm remote control system based on artificial intelligence and robotics technology provided for one embodiment of the present invention.

[0055] In the figure: 10, robotic arm building module; 101, data acquisition module; 102, motion model building module; 20, robotic arm preview module; 201, high-fidelity 3D pre-planning module; 202, instruction conversion module; 203, path optimization module; 30, robotic arm control module. DETAILED DESCRIPTION

[0056] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0057] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0058] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it an embodiment that is mutually exclusive with other embodiments, either individually or selectively.

[0059] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0060] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0061] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0062] Example 1, reference Figure 1 , which is the first embodiment of the present invention, provides a method for remotely controlling a robotic arm based on artificial intelligence and robotics technology, comprising:

[0063] S1: Build the robotic arm in the developed intelligent car lower computer and remotely control the robotic arm through the car behavior control module.

[0064] The present invention is operated in the developed intelligent car system, which consists of a host computer and a slave computer. The host computer is responsible for environmental identification and human-computer interaction, and the slave computer is responsible for behavior control. The host computer uses the trained DeeplabV3+ neural network model to interact with the ChatGPT-based web page through WiFi technology to achieve accurate identification of environmental elements and real-time response to user commands. The slave computer is based on the STM32F746 Discovery development platform, and realizes the communication module, car behavior control module and language interaction module. The cascade PID control strategy optimizes the motor and steering gear control to ensure the accuracy and stability of the car's behavior.

[0065] S2: Build an improved robotic arm through servo motors and control boards, collect the motion parameters of the robotic arm, connect the position relationship between the robotic arm joints and the end effector, and build a robotic arm motion model, including the change in joint angles, the dynamic relationship between joint velocity and acceleration, and the end effector position;

[0066] Furthermore, a high-torque servo motor is used as the driving unit of the robot arm, and a fuzzy control algorithm combined with adaptive control is used to dynamically adjust the output of the motor;

[0067] A camera and force sensor are installed at the front end of the robotic arm to detect object recognition and grasping force, and communicate with the robot's lower computer through an efficient single-chip microcomputer and embedded control board.

[0068] Furthermore, the motion parameters of the robot arm include motion state parameters, environmental parameters and motion modal parameters;

[0069] The motion state parameters include joint rotation angle, joint rotation speed, and joint rotation acceleration;

[0070] Environmental parameters include the weight of the object carried by the end effector, the effect of temperature changes in the working environment on motion performance, and the magnitude of friction between joints;

[0071] The motion modal parameters include movement mode, control mode including autonomous control and remote control, and the initial state of the robot arm before starting to move.

[0072] It should be noted that the relationship between the motion reference frame of the robot arm and the position of the end effector is defined by the feedback control amount and the environmental adaptation amount to feedback the change in joint angle within a specific time period.

[0073]

[0074] in, is the angle of the j-1 joint, the previous joint of the j-th joint, at the previous moment, is the weight coefficient of the control error, is the control error of the jth joint, calculated as the difference between the desired angle and the current angle; is the weight coefficient of environmental adaptation factors, is the environmental parameter influence, caused by load or friction factors; j is the variable index;

[0075] According to the change of joint angle, the dynamic relationship between joint velocity and acceleration is defined:

[0076]

[0077] in, is the velocity of the jth joint at time t, is the time interval, indicating the time period during which the joint displacement occurs; is the velocity of the previous joint j-1, is the influence factor of the previous state on the current speed, controlling the dynamic relationship between the previous and next states;

[0078] Define the end effector position as:

[0079]

[0080] Where P(t) is the position vector of the end effector at time t, The starting position of the robot arm, The length of the jth joint, The adjustment coefficient is used to amplify the impact of environmental factors on the end position; n is the number of joints in the robot arm.

[0081] S3: The robot motion model is input into the high-fidelity 3D pre-planning module for preview and optimization. The RRT algorithm is used to generate potential motion paths. By observing the motion trajectory of the robot, the path planning is adjusted.

[0082] S4: The command transmission mechanism between the upper computer and the lower computer of the smart car is designed as a master-slave mode, with the upper computer as the master control party and the lower computer as the slave control party.

[0083] Furthermore, by inputting the motion model of the robot into the high-fidelity 3D pre-planning module, the calculated parameters in the motion model of the robot are integrated to form a complete state space. At this time, all data are consistent with the real-time feedback of the servo motor and the control board;

[0084] The RRT algorithm is used to generate potential motion paths to ensure that the paths will not collide with obstacles in the virtual environment. The generated paths are executed in the simulation environment. By observing the motion trajectory of the robot arm, the repeatability deviation of the path of the robot arm movement is kept less than 2% in five consecutive simulations. The feasibility and efficiency of the path are evaluated, and the model parameters are adjusted for reverse optimization. When the efficiency of the previewed motion trajectory after optimization is improved by less than 15%, the deep learning algorithm is automatically enabled to iteratively adjust the path planning, and the number of iterations for each adjustment does not exceed 3 times; when the deviation between the actual path and the ideal path exceeds 5 cm, the system will trigger an alarm and require manual adjustment;

[0085] When the user issues a new operation instruction, the host computer should automatically parse and update the motion model through the natural language processing module, re-rehearse, save each rehearsal record, and output the final optimized result to the robot arm control module.

[0086] S5: Input operation instructions through the upper computer, establish a semantic understanding model based on natural language processing to convert the user's natural language instructions into robot arm operation instructions, transmit them to the lower computer through the network, apply the BERT machine learning model for semantic understanding, and the lower computer parses the commands and converts them into motion control instructions, including the rotation angle and movement distance of each axis of the robot arm.

[0087] Furthermore, the user issues natural language commands through voice recognition or text input of the input device, and after preliminary cleaning to remove stop words, an improved word segmentation algorithm is used to identify semantic units, and a dictionary containing robot arm professional terms, verbs, nouns and common phrases is constructed;

[0088] Extract terms from the cleaned text and match them with words in the dictionary. Use a greedy algorithm to match the longest term word by word from left to right. After matching, use natural language processing tools to tag the extracted words with parts of speech, further identify nouns, verbs, and adjectives, and use an improved word segmentation algorithm to build the final operation instructions: Apply BERT's machine learning model for semantic understanding, extract intent and parameters from the instructions, and define new functions ,in, represents the input natural language, 0 is the parsed operation instruction form, and by defining the conversion matrix T, the extracted intent and parameters are mapped to the operation instruction format of the robot arm: , where, C, is the robot arm motion command;

[0089] For example: pre-process and remove noise, unify into "please grab the red cup on the table"; perform preliminary word segmentation, and initially identify as: ["please", "grab", "table", "on", "red", "cup"], after part-of-speech tagging, identify "grab" as a verb, other words as nouns or modifiers, and confirm the spatial relationship between "table" and "cup"; extract operation instructions: the action is "grab", the target is "red cup"; construct it into an instruction format, the final operation command may be "grab (R, B)", "R" represents red, and "B" represents cup.

[0090] The TCP / IP protocol is used to send the constructed operation instructions to the lower computer through the network. The lower computer parses the command and converts it into motion control instructions. The fixed proportional relationship obtained based on the motion model of the robot arm ,in, Represents the dynamic proportional factor between joint i and joint j, quantifying the influence of joint j on the movement of joint i when it rotates to; Represents the static proportional factor between joint i and joint j, that is, the initial value without dynamic adjustment; Indicates the current angle of joint j How does it affect the scale factor when reaching or exceeding 30 degrees? Function of

[0091]

[0092] in, is the current angle of joint j, An adjustment factor to control the degree of nonlinearity of the influence.

[0093] It should also be noted that when joint i needs to be adjusted, it is divided into joint x related to joint i and joint y not related to joint i. When it is a related joint x, the moving distance of the related joint x is adjusted as:

[0094]

[0095] in, represents the moving distance of the i-th joint, represents the basic moving distance of the i-th joint, that is, the initial distance of the current joint without any additional adjustment; Used to indicate whether the angle of joint j reaches 30 degrees. When it exceeds 30 degrees, When it is 1, the adjustment logic is activated; when it does not reach 30 degrees, 0, no adjustment;

[0096] When it is the related joint y, the angle remains the value given by the instruction:

[0097]

[0098] in, Represents the angle change of joint j; the adjusted moving distance and angle are combined to generate motion instructions:

[0099]

[0100] in, is the joint rotation angle, The distance the joint moves, represents the initial distance before adjustment.

[0101] S6: Use reinforcement learning algorithm to optimize and adjust the path planning of the robot arm by adjusting the reward function to find the optimal operation path.

[0102] It should be noted that when the user enters a command, the current status is fed back in real time, and encouraging information is displayed. Based on the feedback, the user is allowed to dynamically adjust the parameters b and c of the reward function and select the optimal operation path:

[0103]

[0104] in, It represents the path with the largest cumulative reward among all possible paths; S represents all possible paths, which is the entire sequence of actions performed by the robot; represents the accumulation from time t=0 to t=T, indicating that the reward is calculated for each time step in the entire path; Indicates the current state The higher the penalty, the farther the current state is from the target. Indicates the current state The higher the reward, the smoother the path.

[0105] S7: The preview and optimization results of the high-fidelity 3D pre-planning module are input into the robot control module for remote control, and the instructions are executed according to the optimal operation path.

[0106] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0107] Embodiment 2, the second embodiment of the present invention, is different from the first two embodiments in that:

[0108] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0109] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0110] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0111] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0112] Example 3, reference Figure 2 , which is the fourth embodiment of the present invention, and provides a robot arm remote control system based on artificial intelligence and robotics technology, including a robot arm building module 10, a robot arm preview module 20, and a robot arm control module 30;

[0113] The robot arm building module 10 includes a data acquisition module 101 and a motion model building module 102. The data acquisition module 101 regularly transmits the collected parameters to the motion model building module 102. The motion model building module 102 analyzes the parameters, builds a motion model of the robot arm, defines the dynamic relationship between the joint angle change, velocity and acceleration, and feeds back the motion model to the data acquisition module 101 for calibration;

[0114] The robot arm preview module 20 includes a high-fidelity 3D pre-planning module 201, an instruction conversion module 202, and a path optimization module 203. When the user issues a work instruction, the instruction conversion module 202 converts the natural language instruction into a robot arm operation instruction, parses it into a motion control instruction, and inputs it into the high-fidelity 3D pre-planning module 201. The motion model is used to generate and optimize the motion path of the robot arm, and the RRT algorithm is used to ensure that the path is safe and effective. The optimization result is then sent back to the path optimization module 203 for deep learning adjustment to improve the working efficiency of the robot arm.

[0115] The robot control module 30 receives the final optimized path and control instructions from the robot preview module 20, controls the movement of the robot in real time, and returns the current state to the robot preview module 20 through a feedback mechanism to further optimize the path.

[0116] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A remote control method for a robotic arm based on artificial intelligence and robotics technology, characterized in that: include, The robotic arm is built in the developed intelligent car lower computer, and the robotic arm is remotely controlled through the car behavior control module; The improved robotic arm is constructed through the servo motor and control board, and the motion parameters of the robotic arm are collected, the position relationship between the robotic arm joints and the end effector is connected, and the robotic arm motion model is constructed, including the change in joint angle, the dynamic relationship between joint velocity and acceleration, and the end effector position; The robot motion model is input into the high-fidelity 3D pre-planning module for preview and optimization, and the RRT algorithm is used to generate potential motion paths. By observing the motion trajectory of the robot, the path planning is adjusted; The command transmission mechanism between the upper computer and the lower computer of the smart car is designed as a master-slave mode, with the upper computer as the master control party and the lower computer as the slave control party; Input operation instructions through the upper computer, establish a semantic understanding model based on natural language processing to convert the user's natural language instructions into robot arm operation instructions, transmit them to the lower computer through the network, apply the BERT machine learning model for semantic understanding, and the lower computer parses the command and converts it into motion control instructions, including the rotation angle and movement distance of each axis of the robot arm; Use reinforcement learning algorithms to optimize and adjust the path planning of the robot arm by adjusting the reward function to find the optimal operation path; Input the preview and optimization results of the high-fidelity 3D pre-planning module into the robotic arm control module for remote control; The high-fidelity 3D pre-planning module performs preview and optimization, including inputting the motion model of the robot arm into the high-fidelity 3D pre-planning module, integrating the calculated parameters in the motion model of the robot arm to form a complete state space, at which point all data are consistent with the real-time feedback of the servo motor and the control board; The RRT algorithm is used to generate potential motion paths to ensure that the paths will not collide with obstacles in the virtual environment. The generated paths are executed in the simulation environment. By observing the motion trajectory of the robot arm, the repeatability deviation of the path of the robot arm movement is kept less than 2% in five consecutive simulations. The feasibility and efficiency of the path are evaluated, and the model parameters are adjusted for reverse optimization. When the efficiency of the previewed motion trajectory after optimization is improved by less than 15%, the deep learning algorithm is automatically enabled to iteratively adjust the path planning, and the number of iterations for each adjustment does not exceed 3 times; when the deviation between the actual path and the ideal path exceeds 5 cm, the system will trigger an alarm and require manual adjustment; When the user issues a new operation instruction, the host computer should automatically parse and update the motion model through the natural language processing module, re-rehearse, save each rehearsal record, and output the final optimized result to the robot arm control module.

2. The method for remotely controlling a robotic arm based on artificial intelligence and robotics technology as claimed in claim 1, characterized in that: The improved mechanical arm includes: using a high-torque servo motor as a driving unit of the mechanical arm, and dynamically adjusting the output of the motor by combining a fuzzy control algorithm with an adaptive control; The front end of the robot arm is equipped with a camera and a force sensor to detect objects and grasping force, and communicate with the robot's lower computer through an efficient single-chip microcomputer and an embedded control board; The motion parameters of the robot arm include motion state parameters, environmental parameters and motion modal parameters; The motion state parameters include joint rotation angle, joint rotation speed, and joint rotation acceleration; The environmental parameters include the weight of the object carried by the end effector, the effect of temperature changes in the working environment on the motion performance, and the magnitude of friction between the joints; The motion mode parameters include movement mode, control mode including autonomous control and remote control, and the initial state of the robot arm before starting to move.

3. The method for remotely controlling a robotic arm based on artificial intelligence and robotics technology as claimed in claim 2, characterized in that: The construction of the robot arm motion model includes defining the relationship between the robot arm motion reference system and the end effector position, and the joint angle change amount in a specific time period is fed back by the feedback control amount and the environmental adaptation amount. : in, is the joint before the jth joint The angle of the joint at the last moment, is the weight coefficient of the control error, is the control error of the jth joint, calculated as the difference between the desired angle and the current angle; is the weight coefficient of environmental adaptation factors, is the environmental parameter influence, caused by load or friction factors; j is the variable index; According to the change of joint angle, the dynamic relationship between joint velocity and acceleration is defined: in, is the velocity of the jth joint at time t, is the time interval, indicating the time period during which the joint displacement occurs; For the previous joint speed, is the influence factor of the previous state on the current speed, controlling the dynamic relationship between the previous and next states; Define the end effector position as: in, The position vector of the end effector at time t, The starting position of the robot arm, The length of the jth joint, The adjustment coefficient is used to amplify the impact of environmental factors on the end position; n is the number of joints in the robot arm.

4. The method for remotely controlling a robotic arm based on artificial intelligence and robotics technology as claimed in claim 3, characterized in that: The semantic understanding model based on natural language processing includes: the user issues natural language instructions through voice recognition or text input of the input device, removes stop words through preliminary cleaning, uses an improved word segmentation algorithm to identify semantic units, and constructs a dictionary containing robot arm professional terms, verbs, nouns and common phrases; Extract terms from the cleaned text and match them with words in the dictionary. Use a greedy algorithm to match the longest term word by word from left to right. After matching, use natural language processing tools to tag the extracted words with parts of speech, further identify nouns, verbs, and adjectives, and use an improved word segmentation algorithm to build the final operation instructions: Apply BERT's machine learning model for semantic understanding, extract intent and parameters from the instructions, and define new functions ,in, is the natural language representing the input, The extracted intentions and parameters are mapped to the operation instruction format of the robot arm by defining the transformation matrix T: , where, C, is the robot arm motion command; The TCP / IP protocol is used to send the constructed operation instructions to the lower computer through the network. The lower computer parses the command and converts it into motion control instructions. The fixed proportional relationship obtained based on the motion model of the robot arm ,in, Represents the dynamic proportional factor between joint i and joint j, quantifying the influence of joint j on the movement of joint i when it rotates to; Represents the static proportional factor between joint i and joint j, that is, the initial value without dynamic adjustment; Indicates the current angle of joint j How does it affect the scale factor when reaching or exceeding 30 degrees? Function of in, is the current angle of joint j, An adjustment factor to control the degree of nonlinearity of the influence.

5. The method for remotely controlling a robotic arm based on artificial intelligence and robotics technology as claimed in claim 4, characterized in that: The semantic understanding model based on natural language processing also includes that when joint i needs to be adjusted, it is distinguished into joints x related to joint i and joints y not related to joint i. When it is a related joint x, the moving distance of the related joint x is adjusted to: in, represents the moving distance of the i-th joint, represents the basic moving distance of the i-th joint, that is, the initial distance of the current joint without any additional adjustment; Used to indicate whether the angle of joint j reaches 30 degrees. When it exceeds 30 degrees, When it is 1, the adjustment logic is activated; when it does not reach 30 degrees, 0, no adjustment; When it is the related joint y, the angle remains the value given by the instruction: in, Represents the angle change of joint j; the adjusted moving distance and angle are combined to generate motion instructions: in, is the joint rotation angle, The distance the joint moves, To represent the initial distance before adjustment, Gives a value to the input command.

6. The method for remotely controlling a robotic arm based on artificial intelligence and robotics technology as claimed in claim 5, characterized in that: The optimal operation path includes providing real-time feedback of the current state when the user inputs a command, and displaying encouraging information. Based on the feedback, the user is allowed to dynamically adjust the parameters b and c of the reward function to select the optimal operation path: in, It represents the path with the largest cumulative reward among all possible paths; Represents all possible paths, which are the entire sequence of actions performed by the robot; Indicates from time arrive The accumulation of , which means that the reward is calculated for each time step in the entire path; Indicates the current state , The higher the penalty, the farther the current state is from the target. Indicates the current state , The smoothness reward, the higher the reward, the smoother the path; The preview and optimization results of the high-fidelity 3D pre-planning module are input into the robotic arm control module for remote control, and the instructions are executed according to the optimal operation path.

7. A system using the method for remotely controlling a robotic arm based on artificial intelligence and robotics technology as claimed in any one of claims 1 to 6, characterized in that: Including robot arm building module, robot arm preview module, and robot arm control module; The robot arm building module includes a data acquisition module and a motion model building module. The data acquisition module regularly transmits the collected parameters to the motion model building module. The motion model building module analyzes the parameters, builds a motion model of the robot arm, defines the dynamic relationship between the joint angle change, velocity and acceleration, and feeds the motion model back to the data acquisition module for calibration; The robot arm preview module includes a high-fidelity 3D pre-planning module, an instruction conversion module and a path optimization module. When the user issues a work instruction, the instruction conversion module converts the natural language instruction into a robot arm operation instruction, parses it into a motion control instruction, inputs it into the high-fidelity 3D pre-planning module, uses the motion model to generate and optimize the motion path of the robot arm, uses the RRT algorithm to ensure the path is safe and effective, and then transmits the optimization result back to the path optimization module for deep learning adjustment to improve the working efficiency of the robot arm; The robot control module receives the final optimized path and control instructions from the robot preview module, controls the movement of the robot in real time, and returns the current state to the robot preview module through a feedback mechanism to further optimize the path.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for remotely controlling a robotic arm based on artificial intelligence and robotics technology described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for remotely controlling a robotic arm based on artificial intelligence and robotics technology described in any one of claims 1 to 6 are implemented.

Images