An intelligent robot control system

Abstract

The application discloses an intelligent mechanical arm control system, and relates to the technical field of intelligent mechanical arms.The system comprises a sensor data acquisition module, a multi-modal data fusion module, an environment and target prediction module, an intelligent decision and control module, and a control execution and feedback module.Through the cooperative working mechanism of the environment and target prediction module and the intelligent decision and control module, the system can accurately predict the future state change of a target object, which makes it possible for the mechanical arm to make a predictive operation, and then a decision optimization algorithm and a multi-objective optimization function are used to plan an optimal motion strategy for the mechanical arm.The system also has real-time feedback and adjustment capabilities, which ensure that the mechanical arm can adapt to changes in a complex dynamic environment autonomously, improve the working precision and efficiency of the mechanical arm, enhance the ability of the mechanical arm to adapt to complex environments and respond to emergencies, and improve production efficiency and safety.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of intelligent mechanical arms, and particularly relates to an intelligent mechanical arm control system. BACKGROUND

[0002] With the advent of the industrial 4.0 era, intelligent manufacturing and industrial automation have become the key force to promote the transformation and upgrading of global manufacturing. As a core execution unit, the performance and intelligent level of a mechanical arm are directly related to production efficiency and product quality. With the increasing complexity and diversification of production tasks, higher requirements are put forward for the sensing ability, decision-making ability and environmental adaptability of the mechanical arm.

[0003] The patent with the application number CN114505856A discloses a mechanical arm intelligent control system based on big data deep learning, which is difficult to handle complex and variable actual working scenes, especially when facing unknown obstacles or dynamic targets, the decision-making efficiency and accuracy are greatly discounted, the real-time changes of environmental information and target states are ignored, resulting in a large deviation between the decision-making results and actual demands, the control strategy cannot be adjusted in time to adapt to new situations, and an effective real-time adjustment mechanism is lacked, so that the accuracy and effectiveness of the control instruction are greatly discounted. Therefore, an intelligent mechanical arm control system is developed. SUMMARY

[0004] The purpose of the application is to make up for the deficiencies of the prior art, and provide an intelligent mechanical arm control system. The system integrates sensor data acquisition, multi-modal data fusion, environment and target prediction, intelligent decision-making and control, and control execution and feedback modules. The intelligent decision-making and control module can receive and preprocess the fused environmental information and prediction results, use a decision optimization algorithm to optimize the control input sequence, plan the motion strategy of the mechanical arm moving chassis and each joint, and generate control instructions. The system also has the ability to adjust the instructions in real time.

[0005] To solve the above technical problems, the application provides the following technical scheme: an intelligent mechanical arm control system, which comprises a sensor data acquisition module, a multi-modal data fusion module, an environment and target prediction module, an intelligent decision-making and control module, and a control execution and feedback module.

[0006] The sensor data acquisition module: a plurality of sensors are deployed on the mechanical arm to construct a multi-dimensional perception network. The original data of the distance, image, joint angle, sound and temperature field of the environment around the mechanical arm are collected in real time at different frequencies and transmitted to the data processing center for storage.

[0007] Multimodal data fusion module: Extracts collected data from the data processing center, performs spatiotemporal calibration and feature extraction, and uses fusion algorithms to fuse feature data from different types of sensors to generate fused environmental information;

[0008] Environment and target prediction module: Based on the fused environmental information and combined with historical data, it analyzes the data from both time and space dimensions. Based on the analysis results, it uses the target state prediction integration model to make short-term predictions of the future state of the target object.

[0009] Intelligent decision and control module: Receives and preprocesses the fused environmental information and prediction results, uses decision optimization algorithm to optimize the control input sequence, plans the motion strategy of the robotic arm's moving chassis and each joint, and generates control commands to send to the control execution and feedback module. At the same time, it receives the actual motion state information of the robotic arm from the control execution and feedback module and uses the robotic arm motion optimization algorithm to adjust the commands in real time.

[0010] Control execution and feedback module: Receives control commands generated by the intelligent decision and control module, drives the robotic arm to move, and collects the actual state data of the robotic arm in real time through sensors during operation, evaluates the target state deviation and environmental interference, and feeds it back to the intelligent decision and control module.

[0011] Furthermore, the sensor used in the sensor data acquisition module is:

[0012] Distance sensor: LiDAR, installed at the center of the top of the robotic arm;

[0013] Visual sensor: A binocular stereo camera is used and mounted on the outside of the forearm;

[0014] Position sensors: High-precision absolute encoders are installed at the rotation axes of each joint;

[0015] Sound sensor: A ring array of MEMS microphones is installed around the upper half of the robotic arm's outer shell;

[0016] Infrared sensor: The FLIRLEPTO 3.5 thermal imaging module is used and is installed on the side of the end effector.

[0017] Furthermore, the multimodal data fusion module employs a fusion algorithm to fuse feature data from different types of sensors. The fusion algorithm formula is as follows: ,in: The fused environmental information vector For quantum state fusion function, To simulate the coupling coefficient of quantum entanglement, For the first Dynamic weights of modal data, the number of modalities participating in fusion, the number of modalities participating in fusion, the index variable.

[0018] Further, the specific steps of analyzing the fused environment information from the time dimension and the space dimension in the environment and target prediction module are:

[0019] Analysis of the time dimension: arrange the fused environment information in chronological order to form time series data of the target object state, and perform smoothing processing, while determining the change trend of position, attitude and temperature, judging the speed change state, detecting whether there is a periodic change rule in the time series, and synchronously performing outlier detection on the time series data;

[0020] Analysis of the spatial dimension: build a three-dimensional coordinate system in the working area of the robot arm, and map the target object and the obstacles in the surrounding environment into the coordinate system, calculate the relative distance and azimuth angle of the target and the surrounding objects according to their spatial coordinate positions in the three-dimensional coordinate system, and extract the features of the spatial structure of the closed area and the motion channel by observing the distribution form of the environmental objects, and evaluate the influence mode and degree of the surrounding environmental factors on the target object attitude and temperature.

[0021] Further, the environment and target prediction module performs short-term prediction on the future state of the target object according to the analysis results through a target state prediction integration model, and the calculation formula of the target state prediction integration model is: wherein, is the state vector of the predicted target object at the next time point, is the fusion weight of the time dimension and the space dimension prediction results, is the state predicted by the time series model, which is obtained by training the historical time series data, the number of spatially related objects, the number of spatially related objects, the weight of the th spatially related object, k is the index variable, k is the number of surrounding objects included in the analysis. Further, the intelligent decision and control module optimizes the control input sequence of the robot arm according to a decision optimization algorithm, and the formula of the decision optimization algorithm is:

[0022] wherein, is the control input sequence of the robot arm, is the motion energy consumption function, ​​a distance penalty term for the robot arm and the obstacle, a variation of adjacent control inputs, and a weight coefficient dynamically adjusted according to the obstacle density and the ambient temperature change rate.

[0023] Further, the specific steps of planning the robot arm moving chassis and joint motion strategy in the intelligent decision and control module are as follows: analyzing the received fusion environment information and prediction results, determining the type, priority, accuracy requirement of the current robot arm task, and the position, posture, motion trend of the target object and the surrounding environment obstacle distribution information, constructing a multi-objective optimization function with the task completion time, energy consumption, motion stability, and safety as the core targets according to the real-time distance between the target object and the robot arm, the target prediction motion speed and acceleration parameters, setting the corresponding weight coefficient for each target according to the actual requirements of the task, combining the kinematic and dynamic constraints of the robot arm to form a multi-objective optimization model, and solving the multi-objective optimization model, and through continuous iterative calculation, obtaining the control parameters of the moving speed, steering angle of the robot arm moving chassis, and the rotation angle, angular velocity, and angular acceleration of each joint.

[0024] Further, the intelligent decision and control module constructs a multi-objective optimization function with the task completion time, energy consumption, motion stability, and safety as the core targets. Specifically, let be the optimization objective function, be the task completion time, be the reference completion time, be the energy consumption, be the reference energy consumption, be the motion stability index, be the safety evaluation index, then the multi-objective optimization function formula is: ; : ; ; ; ; wherein, , , , are the weight coefficients of the corresponding targets, and is dynamically adjusted according to the task type, represents the kinematic constraint function of the robot arm, represents the dynamic constraint function of the robot arm, is the joint angle vector of the robot arm, is the joint angular velocity vector, is the joint angular acceleration vector, is a real-time distance between the target object and the robot arm, which is measured by sensor data, is a maximum allowed distance value, is a target predicted speed, which is predicted by a prediction model based on the motion speed of the robot arm, is a maximum allowed speed value, is a target predicted acceleration, which is calculated by a prediction model based on the motion state data of the robot arm, is a maximum allowed acceleration value.

[0025] Further, the intelligent decision and control module uses a robot arm motion optimization algorithm to adjust the instruction in real time, and the robot arm motion optimization algorithm formula is: , wherein: is the target state at time t, the robot arm control instruction adjusted in real time by the optimization algorithm, is the received control instruction of the intelligent decision and control module, is a deviation vector between the actual state and the target state of the robot arm, is an environmental disturbance intensity estimation value, is a feedback signal, is time.

[0026] Compared with the prior art, the intelligent robot arm control system has the following beneficial effects:

[0027] First, the present application can accurately predict the future state change of the target object through the cooperative working mechanism of the environment and target prediction module and the intelligent decision and control module, which provides the possibility for the predictive operation of the robot arm, and then uses the decision optimization algorithm and the multi-objective optimization function to plan the optimal motion strategy for the robot arm. The system also has real-time feedback and adjustment capabilities, ensuring that the robot arm can adapt to changes in a complex dynamic environment, improving the working precision and efficiency of the robot arm, and enhancing its ability to adapt to complex environments and respond to unexpected situations, improving production efficiency and safety.

[0028] Second, the present application provides key support for the intelligent upgrading of the system through the control execution and feedback module. The sensor deployed at each key position of the robot arm can collect the motion state data of the robot arm in real time, which is used to feedback the execution effect to the intelligent decision and control module, so that the robot arm can be upgraded from a passive execution instruction tool to an intelligent agent with autonomous state monitoring capability, effectively enhancing the autonomous decision-making capability of the robot arm and laying a foundation for stable execution of complex tasks.

[0029] Additional advantages, objects, and features of the application will be apparent to those skilled in the art upon examination of the following detailed description, it being understood that each BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments or the prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0031] Figure 1 It is a structural schematic diagram of an intelligent mechanical arm control system.

[0032] Figure 2 It is a flow chart of an intelligent mechanical arm control system. DETAILED DESCRIPTION

[0033] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined application purpose, the specific embodiments, structures, features and effects according to the present application will be described in detail below with reference to the drawings and preferred embodiments.

[0034] Embodiment one

[0035] In an automated warehouse center, the mechanical arm needs to accurately grasp, carry and sort packages of different sizes and weights, while avoiding dynamic obstacles and adapting to complex lighting and temperature environments in the warehouse.

[0036] Sensor data acquisition module: 20Hz laser radar (distance sensor) is installed at the top center of the mechanical arm, which scans the spacing between sorting shelves and the position of packages in real time; 20Hz binocular stereo camera (visual sensor) is deployed on the outside of the forearm, which identifies the barcode and three-dimensional size of the package; 100Hz high-precision absolute encoder (position sensor) is configured on each joint rotation axis to monitor the grasping posture of the mechanical arm; 50Hz ring MEMS microphone array (sound sensor) is arranged around the mechanical arm shell to detect abnormal sound of the conveying belt; 50Hz FLIRLepton3.5 thermal imaging module (infrared sensor) is installed on the side of the end effector to sense temperature abnormalities on the surface of the package; The raw data (distance, image, joint angle, sound, temperature) collected are transmitted synchronously to the data processing center for storage.

[0037] Multi-modal data fusion module: extract sensor data from the data processing center, such as Figure 1As shown, the feature data of different types of sensors are fused by a fusion algorithm, and the formula of the fusion algorithm is: wherein: is the fused environmental information vector (such as the comprehensive data of package position, size, and conveyor belt speed), is a quantum state fusion function, is a coupling coefficient simulating quantum entanglement, is the dynamic weight of the th modal data (such as the weight of laser radar distance data being higher than that of sound data), is the original data of the th modal, is an index variable, is the number of modes participating in fusion; the quantum state fusion function simulates the characteristics of quantum entanglement, enhances the spatial correlation of different modal data, and dynamically adjusts the weights. After data fusion, the environmental information including package position, size, barcode information, conveyor belt speed, and environmental safety is generated.

[0038] The environmental and target prediction module: based on the fused environmental information, combined with historical data, through time dimension analysis, the data of package position and conveyor belt speed are arranged, the position and speed change trend are determined through smoothing processing, and the periodic law (such as the start-stop period of the conveyor belt) and abnormal value (such as the emergency stop event) are detected; through spatial dimension analysis, a three-dimensional coordinate system is established in the sorting area, the positions of the package, the conveyor belt, and the sorting port are mapped, the relative distance and azimuth angle are calculated, the package accumulation density in the channel is analyzed, and according to the results of time dimension and spatial dimension analysis, the state of the package at the next time point is predicted through a target state prediction integration model, and the calculation formula of the target state prediction integration model is: wherein: is the state vector (position, attitude) of the predicted target object at the next time point, is the fusion weight of the prediction results of time dimension and spatial dimension (such as the proportion of time dimension being 60%), is the prediction value of the time series model, which is obtained by training the historical time series data, is the correction value of the th spatially related object to the target state, which is obtained by statistically analyzing the change law of the target state under the spatial relationship in the historical data, is the weight of the k th spatially related object, k is an index variable, is the number of surrounding objects included in the analysis (such as = 3).

[0039] Intelligent decision and control module: after receiving and preprocessing the fused environmental information and prediction results, the control input sequence is optimized through a decision optimization algorithm, and the formula of the decision optimization algorithm is: wherein: is the mechanical arm control input sequence (such as joint angle, movement speed), is the motion energy consumption function, is the distance penalty term between the mechanical arm and the obstacle (such as increasing the penalty when the distance to the conveyor belt edge is too close), is the change amount of adjacent control input, and are weight coefficients dynamically adjusted according to the obstacle density and the change rate of the environmental temperature (automatically adjusted according to the obstacle density, such as increasing when the packages are stacked), the control input sequence is output to plan the movement strategy of the mechanical arm chassis and each joint;

[0040] At the same time, a multi-objective optimization function is constructed with the core objectives of task completion time, energy consumption, motion smoothness, and safety, and according to the actual requirements of the task, the corresponding weight coefficients are set for each objective, combined with the kinematic and dynamic constraints of the mechanical arm, a multi-objective optimization model is formed, and the multi-objective optimization model is solved, specifically, let be the optimization objective function, be the task completion time, be the reference completion time, be the energy consumption, be the reference energy consumption, be the motion smoothness index, be the safety evaluation index, then the formula of the multi-objective optimization function is: wherein, , , , are the weight coefficients of the corresponding objectives respectively, and are dynamically adjusted according to the task type, represents the kinematic constraint function of the mechanical arm, represents the dynamic constraint function of the mechanical arm, is the joint angle vector of the mechanical arm, is the joint angular velocity vector, is the joint angular acceleration vector, is the real-time distance between the target object and the mechanical arm, which is measured by the sensor data, is the maximum allowed distance value, ​​​​​​The target predicted speed is obtained by predicting the movement speed of the mechanical arm through a prediction model, The allowed maximum speed value, The target predicted acceleration is calculated by the prediction model according to the data of the movement state of the mechanical arm, The allowed maximum acceleration value is obtained by continuously iterating the calculation of the moving speed of the chassis of the mechanical arm, the steering angle, and the control parameters of the rotation angle, angular velocity, and angular acceleration of each joint.

[0041] The control execution and feedback module: through the generated control instruction, the mechanical arm is driven to perform the grabbing action, and through the sensor, the joint angle, jaw pressure and other data are collected in real time, the deviation from the target state is calculated, the interference such as the vibration of the conveying belt is evaluated, the deviation information is fed back to the intelligent decision and control module, and the mechanical arm movement optimization algorithm is triggered to adjust the next action, and the formula of the mechanical arm movement optimization algorithm is: Wherein: is the time at which The mechanical arm control instruction after real-time adjustment by the optimization algorithm, The control instruction received by the intelligent decision and control module, The deviation vector of the actual state of the mechanical arm and the target state, The estimated value of the environmental disturbance intensity, The feedback signal, The time.

[0042] As described above, in the logistics sorting scene, the scheme can accurately perceive the position, size and state of the conveying belt through multi-modal sensor fusion and dynamic prediction model, significantly improve the sorting efficiency and accuracy, the intelligent decision module reduces the mechanical arm idle and collision risk by optimizing energy consumption and obstacle avoidance strategy, and the closed-loop feedback mechanism real-time corrects the grabbing deviation, adapts to the complex environment such as the vibration of the conveying belt and the accumulation of the package, improves the sorting efficiency and supports 24-hour continuous operation, and is suitable for high-load scenes such as e-commerce promotion and warehouse logistics.

[0043] Embodiment two

[0044] For fine tissue separation and hemostasis operation in laparoscopic surgery, the mechanical arm is required to have sub-millimeter positioning accuracy, real-time obstacle avoidance (avoiding important blood vessels) and heat damage monitoring capability, while balancing the operation speed and tissue protection demand.

[0045] Sensor data acquisition module: such as Figure 2As shown, a 20Hz miniature laser radar is installed at the top center of the mechanical arm to scan the organ profile and instrument spacing within the surgical field; a 20Hz medical binocular endoscope (visual sensor) is integrated on the outside of the forearm to obtain high-definition blood vessel and tissue images of the surgical site in real time; a 100Hz high-precision encoder is configured at each joint to monitor the micron-level displacement of the surgical instrument of the mechanical arm (with a precision of ±0.01mm); a 50Hz medical microphone array is arranged on the upper part of the mechanical arm shell to collect the operating sound characteristics of instruments such as ultrasonic knives and electric knives; a 50Hz infrared thermal imaging module is installed on the side of the end surgical instrument to monitor the temperature changes of the tissue (with a resolution of 0.1℃), and the raw data collected (organ position, tissue image, joint displacement, instrument sound, and tissue temperature) are stored synchronously to the data processing center.

[0046] Multi-modal data fusion module: organ profile data from the laser radar, blood vessel image features from the endoscope, instrument position from the encoder, ultrasonic knife frequency signal from the microphone, and thermal imaging temperature field are extracted from the data processing center, and fusion information is generated through a fusion algorithm formula, which is: .

[0047] Environment and target prediction module: based on the fused environmental information, through time dimension analysis, the instrument motion trajectory (such as electric knife moving speed) and tissue temperature change data are arranged in time sequence, the next time point position of the instrument is predicted through smoothing processing (error ≤0.2mm), the temperature rise rate is analyzed (threshold set to 0.5℃ / s, exceeding which triggers an early warning), and abnormal shaking of the mechanical arm is detected (such as acceleration mutation when colliding with the operating table); through spatial dimension analysis, a three-dimensional coordinate system is established at the patient's surgical site, mapping structures such as liver, blood vessels, and tumors, the relative distance between the instrument and the hepatic portal vein is calculated (a safety distance of ≥2mm needs to be maintained) through the spatial analysis step of claim 4, and the organ displacement amplitude caused by respiratory motion (such as liver fluctuation of 1.5mm up and down) is evaluated; combined with the analysis results, the instrument influence is predicted through a target state prediction integration model, and the calculation formula of the target state prediction integration model is: .

[0048] Intelligent decision and control module: receiving the fused blood vessel position, tissue temperature, and other information and prediction results (such as the electric knife approaching the blood vessel in 0.5 seconds), using a decision optimization algorithm to optimize the control instructions of the mechanical arm, the decision optimization algorithm formula is: , planning the movement of the mechanical arm chassis and the motion strategy of each joint, and constructing a multi-objective optimization function with task completion time, energy consumption, motion stability, and safety as the core targets, setting appropriate weight coefficients for each target according to the actual requirements of the task, combining the constraints of the kinematics and dynamics of the mechanical arm, forming a multi-objective optimization model, and the formula of the multi-objective optimization function is: ; : ; ; ; ; , through continuous iteration calculation, the moving speed of the mechanical arm moving chassis, the steering angle, and the control parameters of the rotation angle, angular velocity and angular acceleration of each joint are obtained.

[0049] The control execution and feedback module: through the generated control instruction, the mechanical arm is driven to perform separation and hemostasis operation, and through the sensor, the instrument position (error ≤0.05mm), tissue contact force (resolution 0.01N) and other data are collected in real time, the deviation of the path is calculated and planned (such as the temperature of the electric knife burning area exceeding the safety threshold 1℃), the interference strength of respiratory motion is evaluated, the deviation information is fed back to the intelligent decision module, the mechanical arm motion optimization algorithm is triggered to adjust the next action, and the formula of the mechanical arm motion optimization algorithm is: .

[0050] In summary, in the medical operation scene, the scheme realizes the accurate modeling and real-time monitoring of the blood vessels and tissues in the operation area through the sub-millimeter level sensor accuracy and quantum state data fusion, the time and space prediction model warns the contact risk of the instrument and the key structure in advance 0.5 seconds, the operation error is controlled within 0.1mm by combining the multi-objective optimization algorithm, the risk of electric knife thermal damage is reduced by 40% through dynamic weight adjustment, and the physiological motion compensation mechanism effectively offsets the organ displacement caused by patient breathing (compensation accuracy 0.05mm), which improves the stability of laparoscopic surgery, especially suitable for high-precision operations such as liver and gallbladder blood vessel separation and microsuture, reduces the amount of intraoperative bleeding and reduces the incidence of postoperative complications.

[0051] The above is only a preferred embodiment of the present application, and is not intended to limit the present application in any form. Although the present application has been disclosed as above with a preferred embodiment, it is not intended to limit the present application. Any person skilled in the art can make some changes or modifications to the above disclosed technical content without departing from the scope of the technical solution of the present application, and any equivalent embodiments with equivalent changes are equivalent. Any modification, change, equivalent change and modification of the above embodiments made according to the technical essence of the present application are still within the scope of the technical solution of the present application.

Claims

1. An intelligent robot control system, characterized by, The system comprises: A sensor data acquisition module: a plurality of sensors are deployed on the robot arm to construct a multi-dimensional perception network, to collect raw data of distance, image, joint angle, sound and temperature field of the surrounding environment of the robot arm in real time at different frequencies, and transmit the data to the data processing center for storage; Multi-modal data fusion module: extract the collected data from the data processing center, carry out space-time calibration and feature extraction, use fusion algorithm to fuse the feature data of different types of sensors, generate fused environment information, and the fusion algorithm formula is: Wherein: is the fused environment information vector, is the quantum state fusion function, is the coupling coefficient simulating quantum entanglement, is the dynamic weight of the first modal data, is the original data of the first modal, is the number of modes participating in fusion, is the index variable; An environment and target prediction module: based on the fused environment information, combined with historical data, respectively from the time dimension and the spatial dimension for analysis, according to the analysis results through the target state prediction integration model to predict the future state of the target object, its target state prediction integration model calculation formula is: Wherein, is the state vector of the predicted target object at the next time point, is the fusion weight of the time dimension and the spatial dimension prediction result, is the state predicted by the time series model, which is obtained by training the historical time series data, is the correction value of the target state of the th spatial related object, which is obtained by statistically analyzing the change rule of the target state under the spatial relationship in the historical data, is the weight of the th k spatial related object, k is the index variable, is the number of surrounding objects included in the analysis; The intelligent decision and control module receives and pre-processes the fused environmental information and prediction results, optimizes the control input sequence using a decision optimization algorithm, plans the moving base and joint motion strategy of the robot arm, and generates control instructions to send to the control execution and feedback module. The decision optimization algorithm formula is: wherein, is the control input sequence of the robot arm, is the motion energy consumption function, is the distance penalty term between the robot arm and the obstacle, is the change amount of adjacent control input, and are weight coefficients dynamically adjusted according to the obstacle density and the environmental temperature change rate; at the same time, the actual motion state information of the robot arm fed back by the control execution and feedback module is received, and the robot arm motion optimization algorithm is used to adjust the instructions in real time. A control execution and feedback module: receiving the control instructions generated in the intelligent decision and control module, driving the robot arm to act, in the running process, collecting the actual state data of the robot arm in real time through the sensors, evaluating the target state deviation and environmental disturbance, and feeding back to the intelligent decision and control module.

2. The intelligent robotic arm control system of claim 1, wherein, The sensors used in the sensor data acquisition module are: Distance sensor: laser radar installed at the top center of the robot arm; Visual sensor: binocular stereo camera installed on the outside of the forearm; Position sensor: high-precision absolute encoder installed at the rotating shaft of each joint; Sound sensor: ring array composed of MEMS microphones installed around the upper half of the robot arm shell; Infrared sensor: FLIR Lepton 3.5 thermal imaging module installed on the side of the end effector.

3. The intelligent robotic arm control system of claim 1, wherein, The specific steps of analyzing the fused environmental information from the time dimension and the space dimension in the environment and target prediction module are: Time dimension analysis: arrange the fused environmental information in chronological order to form time series data of the target object state, and perform smoothing processing, while determining the change trend of position, attitude and temperature, judging the speed change state, detecting whether there is a periodic change rule in the time series, and simultaneously performing outlier detection on the time series data; Spatial dimension analysis: build a three-dimensional coordinate system in the working area of the robot arm, and map the target object and the obstacles in the surrounding environment into the coordinate system, calculate the relative distance and azimuth angle of the target and the surrounding objects according to the spatial coordinate positions of the target object and the obstacles in the surrounding environment in the three-dimensional coordinate system, and extract the spatial structure features of the closed area and the motion channel by observing the distribution form of the environmental objects, to evaluate the influence mode and degree of the surrounding environmental factors on the attitude and temperature of the target object.

4. The intelligent robotic arm control system of claim 1, wherein, The specific steps of planning the movement of the robot arm chassis and the joint motion strategy in the intelligent decision and control module are: analyzing the received fused environmental information and prediction results, determining the type, priority, accuracy requirement of the current robot arm execution task, and the position, attitude, motion trend of the target object and the distribution information of the surrounding environmental obstacles, constructing a multi-objective optimization function with task completion time, energy consumption, motion stability, safety as the core target according to the current real-time distance of the target object and the robot arm, the parameters of the target prediction motion speed and acceleration, setting the corresponding weight coefficients for each target according to the actual requirements of the task, combining the kinematic and dynamic constraints of the robot arm to form a multi-objective optimization model, and solving the multi-objective optimization model, through continuous iterative calculation, obtaining the control parameters of the running speed, steering angle of the robot arm chassis, and the rotation angle, angular velocity, angular acceleration of each joint.

5. The intelligent robotic arm control system of claim 4, wherein, The intelligent decision and control module constructs a multi-objective optimization function with task completion time, energy consumption, motion stability, and safety as core targets. Specifically, let be the optimization objective function, be the task completion time, be the reference completion time, be the energy consumption, be the reference energy consumption, be the motion stability index, be the safety evaluation index, then the multi-objective optimization function formula is: ; : ; ; ; ; wherein, , , , are the weight coefficients of the corresponding targets, and , the weights are dynamically adjusted according to the task type, represents the kinematics constraint function of the robot arm, represents the dynamics constraint function of the robot arm, is the joint angle vector of the robot arm, is the joint angular velocity vector, is the joint angular acceleration vector, is the real-time distance between the target object and the robot arm, which is measured by sensor data, is the allowed maximum distance value, is the target prediction speed, which is obtained by predicting the motion speed of the robot arm using a prediction model, is the allowed maximum speed value, is the target prediction acceleration, which is calculated based on the data of the motion state of the robot arm using a prediction model, is the allowed maximum acceleration value.

6. The intelligent robotic arm control system of claim 1, wherein, The intelligent decision and control module uses a mechanical arm motion optimization algorithm to adjust instructions in real time, and the mechanical arm motion optimization algorithm formula is: Wherein: is the control instruction of the intelligent decision and control module at time The mechanical arm control instruction adjusted in real time through the optimization algorithm, is the received control instruction of the intelligent decision and control module, is the deviation vector of the actual state and the target state of the mechanical arm, is the estimated value of the environmental disturbance intensity, is the feedback signal, is the time.

Images