Adaptive control method and equipment of mechanical booster arm based on multi-feature monitoring

Abstract

The present application provides a method and device for adaptively controlling a mechanical power-assisting arm based on multi-feature monitoring, and relates to the technical field of mechanical arm control. The method includes: obtaining the gripping preparation stage; collecting operating parameters, connecting a data storage unit to obtain work records; constructing a twin network model for simulation; optimizing motion trajectories and grasping strategies to obtain alternative sequences; and determining the optimal combination for control optimization. The present application solves the technical problem that the operation process of a mechanical power-assisting arm relies on fixed programming or simple feedback control, which makes it difficult to adapt to complex and changing working environments and task requirements, resulting in low grasping success rate and stability. It realizes comprehensive monitoring of the mechanical power-assisting arm in the gripping preparation stage, uses a twin network model for path planning preview simulation, improves grasping accuracy, and dynamically adjusts the control strategy according to real-time operating parameters, quickly adapts to different working conditions and environmental changes, and improves the operation success rate and grasping stability.

Description

Technical Field

[0001] The present application relates to the technical field of mechanical power-assist arm control, and in particular to a mechanical power-assist arm adaptive control method and device based on multi-feature monitoring. Background Art

[0002] With the rapid development of industrial automation and intelligent technology, mechanical power-assist arms are increasingly used in various fields. However, the mechanical power-assist arms in the existing technology are not flexible enough in changing environments and have weak adaptability, which often leads to poor performance when facing complex tasks. At the same time, the optimization efficiency of the motion trajectory and grasping strategy of the mechanical power-assist arms is low, and the accuracy needs to be improved, which directly affects the efficiency and stability of the work. Summary of the invention

[0003] The purpose of this application is to provide an adaptive control method and device for a mechanical power-assisting arm based on multi-feature monitoring, so as to solve the technical problem that the operation process of the mechanical power-assisting arm relies on fixed programming or simple feedback control, is difficult to adapt to complex and changing working environments and task requirements, and leads to low grasping success rate and stability.

[0004] In view of the above problems, the present application provides a method and device for adaptively controlling a mechanical power arm based on multi-feature monitoring.

[0005] In the first aspect, the present application provides a method for adaptive control of a mechanical power-assisting arm based on multi-feature monitoring, and the method is implemented by an adaptive control device for a mechanical power-assisting arm based on multi-feature monitoring, wherein the method comprises: obtaining a gripping preparation stage of the mechanical power-assisting arm, the gripping preparation stage comprising a self-checking stage for homing the mechanical arm, a parameter setting stage, and a target object identification and positioning stage; performing feature synchronization monitoring on the mechanical power-assisting arm, collecting real-time operating parameters of the mechanical power-assisting arm, connecting a data storage unit of the mechanical power-assisting arm and obtaining handling work data records, wherein the handling work data records comprise position data, motion trajectory data, speed data, acceleration data, and torque data; comparing the mechanical power-assisting arm, and constructing a twin network model of the mechanical power-assisting arm through the handling work data records, wherein the mechanical power-assisting arm comprises a base component group, a bracket component group, and a joint component group. A component group, a robot arm body, and an end-execution component group, wherein the joint component group includes bearings, gears, and a transmission mechanism; in the grasping preparation stage, based on the twin network model, parameter configuration is performed through the real-time operating parameters, and after completing the parameter configuration, a path planning preview simulation is performed to obtain a simulation result; based on the simulation result, the motion trajectory of the mechanical power-assisting arm is optimized, and a sequence of alternative motion trajectories is obtained in the order of grasping time consumption from small to large; based on the simulation result, the grasping strategy of the mechanical power-assisting arm is optimized, and a sequence of alternative grasping strategies is obtained in the order of grasping success rate from large to small; based on the alternative motion trajectory sequence and the alternative grasping strategy sequence, a machine learning algorithm is used to determine the optimal motion trajectory-grasping strategy combination through the real-time operating parameters and perform regulation and optimization.

[0006] In the second aspect, the present application also provides a mechanical power-assisting arm adaptive control device based on multi-feature monitoring, which is used to execute the mechanical power-assisting arm adaptive control method based on multi-feature monitoring as described in the first aspect, wherein the device includes: a preparation stage acquisition module: the preparation stage acquisition module is used to acquire the gripping preparation stage of the mechanical power-assisting arm, and the gripping preparation stage includes the mechanical arm homing self-inspection stage, the parameter setting stage, and the target object identification and positioning stage; a parameter acquisition module: the parameter acquisition module is used to perform feature synchronization monitoring on the mechanical power-assisting arm, collect the real-time operating parameters of the mechanical power-assisting arm, connect the data storage unit of the mechanical power-assisting arm and obtain the handling work data record, and the handling work data record includes position data, motion trajectory data, speed data, acceleration data, and torque data; a model construction module: the model construction module is used to compare the mechanical power-assisting arm, and through the handling work data record, construct a twin network model of the mechanical power-assisting arm, and the mechanical power-assisting arm includes a base component group, a bracket component group, a joint component group, A robotic arm body and an end-actuator component group, wherein the joint component group includes bearings, gears and transmission mechanisms; a simulation module: the simulation module is used to configure parameters based on the twin network model and the real-time operating parameters during the grasping preparation stage, and perform path planning preview simulation after completing the parameter configuration to obtain simulation results; a trajectory optimization module: the trajectory optimization module is used to optimize the motion trajectory of the robotic power-assisting arm based on the simulation results, and obtain a sequence of alternative motion trajectories in the order of grasping time consumption from small to large; a strategy optimization module: the strategy optimization module is used to optimize the grasping strategy of the robotic power-assisting arm based on the simulation results, and obtain a sequence of alternative grasping strategies in the order of grasping success rate from large to small; a control optimization module, the control optimization module is used to determine the optimal motion trajectory-grasping strategy combination and perform control optimization based on the alternative motion trajectory sequence and the alternative grasping strategy sequence using a machine learning algorithm and the real-time operating parameters.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] The above-mentioned adaptive control method of the mechanical power arm based on multi-feature monitoring first determines the gripping preparation stage of the mechanical power arm, which includes the homing self-check of the mechanical power arm, parameter setting and target object identification and positioning. In this stage, multiple key features of the mechanical power arm are monitored synchronously, its operating parameters are collected in real time, and the data records of the previous handling work are obtained. These data record the position, motion trajectory, speed, acceleration and torque of the mechanical power arm in detail. Subsequently, a twin network model corresponding to the mechanical power arm is constructed using these actual operating data. This model contains various components of the mechanical power arm, such as the base, bracket, joint and end effector. Afterwards, based on this twin network model, the parameters are configured using real-time operating parameters, and a preview simulation of path planning is performed. Through simulation, the simulation results are obtained, and based on these results, the motion trajectory of the mechanical power arm is optimized, and a series of alternative motion trajectories sorted by grasping time consumption are generated. At the same time, the grasping strategy is also optimized to obtain alternative strategies sorted by grasping success rate. Then, combining these alternative motion trajectories and grasping strategies, the optimal combination of motion trajectory and grasping strategy was determined based on the real-time operating parameters using machine learning algorithms, and regulation optimization was performed to ensure that the robotic arm can complete the grasping task efficiently and accurately.

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0011] Figure 1 This is a flow chart of the method for adaptively controlling a mechanical power-assisting arm based on multi-feature monitoring in this application;

[0012] Figure 2 This is a schematic diagram of the structure of the adaptive control device of the mechanical power arm based on multi-feature monitoring in this application.

[0013] Explanation of the accompanying drawings: preparation phase acquisition module 1, parameter collection module 2, model building module 3, simulation module 4, trajectory optimization module 5, strategy optimization module 6, control optimization module 7. DETAILED DESCRIPTION

[0014] The present application provides an adaptive control method and equipment for a mechanical power-assisting arm based on multi-feature monitoring, thereby solving the technical problem that the operation process of the mechanical power-assisting arm relies on fixed programming or simple feedback control, which is difficult to adapt to complex and changeable working environments and task requirements, resulting in low grasping success rate and stability. The application realizes comprehensive monitoring of the mechanical power-assisting arm in the grasping preparation stage, uses the twin network model to perform path planning preview simulation to improve grasping accuracy, and dynamically adjusts the control strategy according to real-time operating parameters to quickly adapt to different working conditions and environmental changes, thereby improving the operation success rate and grasping stability.

[0015] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0016] Embodiment 1

[0017] Please see attached Figure 1 The present application provides a method for adaptively controlling a mechanical power-assisting arm based on multi-feature monitoring, wherein the method is applied to an adaptively controlling device for a mechanical power-assisting arm based on multi-feature monitoring, and the method specifically includes the following contents:

[0018] Acquire the gripping preparation stage of the mechanical power-assisting arm, wherein the gripping preparation stage includes a mechanical arm homing self-checking stage, a parameter setting stage, and a target object identification and positioning stage;

[0019] The complexity and variability of the industrial production environment place higher demands on the performance and stability of the mechanical power arm. The use of multi-feature monitoring can collect a variety of operating parameters and status information of the mechanical power arm in real time, such as position, speed, acceleration, torque, etc., so as to have a more comprehensive understanding of the working status of the mechanical power arm. This enables the mechanical power arm to more accurately identify and handle various complex situations, and improve its adaptability and stability in a changing environment.

[0020] In the embodiment of the present application, determining the gripping preparation stage of the mechanical power-assisting arm is a crucial process, which ensures that the mechanical power-assisting arm is in the best state before performing the grasping task. This process includes the mechanical arm homing self-checking stage, the parameter setting stage, and the target object identification and positioning stage. Among them, in the mechanical arm homing self-checking stage, the mechanical power-assisting arm will return to the initial position and perform a series of self-checking operations. This includes checking whether the various components of the mechanical power-assisting arm are intact and whether the various sensors are working properly. Through the homing self-check, it can be ensured that the mechanical power-assisting arm is in a good state when preparing to perform the grasping task. In the parameter setting stage, the motion parameters of the mechanical power-assisting arm are set according to the specific grasping task requirements. These parameters include the motion speed, acceleration, grasping force, etc. of the mechanical power-assisting arm. Through reasonable parameter setting, the mechanical power-assisting arm can better adapt to the needs of different tasks and improve the grasping efficiency and accuracy. In the target object identification and positioning stage, the mechanical power-assisting arm uses sensors and visual systems to identify and locate the target object. By obtaining information such as the position, shape and posture of the target object, the mechanical power-assisting arm can accurately determine the grasping point and grasping method, and provide accurate guidance for subsequent grasping operations. In summary, by obtaining the gripping preparation stage of the robotic arm, it can be ensured that the robotic arm is in the best condition before performing the gripping task, and accurate guidance and guarantee are provided for subsequent gripping operations.

[0021] Performing feature synchronous monitoring on the mechanical power-assisting arm, collecting real-time operating parameters of the mechanical power-assisting arm, connecting to a data storage unit of the mechanical power-assisting arm and acquiring handling work data records, wherein the handling work data records include position data, motion trajectory data, speed data, acceleration data, and torque data;

[0022] In one embodiment, the system terminal synchronously obtains the real-time operating parameters of the mechanical power-assisting arm, such as position, motion trajectory, speed, acceleration and torque, through the monitoring equipment. These data can intuitively show the dynamic performance of the mechanical power-assisting arm at work. At the same time, the data storage unit of the mechanical power-assisting arm is also connected to obtain the previous handling work data records. These data records record in detail various information of the mechanical power-assisting arm in past work, including the position, motion trajectory, speed, acceleration, torque and other data during each handling. These data not only help the system terminal to detect potential problems in a timely manner, but also provide strong support for optimizing the motion trajectory and grasping strategy of the mechanical power-assisting arm.

[0023] By comparing the mechanical power-assisting arm, a twin network model of the mechanical power-assisting arm is constructed through the handling work data record, wherein the mechanical power-assisting arm includes a base component group, a bracket component group, a joint component group, a mechanical arm body, and an end-actuator component group, wherein the joint component group includes bearings, gears, and a transmission mechanism;

[0024] In one embodiment, the system terminal compares the actual structure and function of the mechanical power arm and uses the acquired handling work data records to build a twin network model corresponding to it. This model is a virtual reproduction of the mechanical power arm, which includes all the main components of the mechanical power arm, such as the base, bracket, joint and end effector. Among them, key components such as bearings, gears and transmission mechanisms in the joint component group are also accurately simulated in the model. Specifically, the system terminal analyzes the structure of the mechanical power arm in detail and divides it into a base component group, a bracket component group, a joint component group, a mechanical arm body and an end effector component group according to its components. For each component group, accurate three-dimensional modeling is performed to ensure that the model can accurately reflect the physical properties and geometric shape of the mechanical power arm. Subsequently, key components such as bearings, gears and transmission mechanisms in the joint component group are modeled in detail. The motion state and interaction of these components are crucial to the overall performance of the mechanical power arm, so it is necessary to ensure that their working principles and dynamic characteristics are accurately simulated in the model. Afterwards, the system terminal designs the network structure according to the structure of the mechanical power arm and the handling work data records. This includes determining the number of nodes in the input layer, hidden layer, and output layer, selecting activation functions and loss functions, and setting optimization algorithms. The model parameters are then initialized using the handling work data records. These parameters include network weights and bias terms, which will affect the output and performance of the model. By adjusting these parameters, the model can better fit the actual working state of the mechanical power arm. After completing the model construction and parameter configuration, the system terminal performs model verification and debugging. The output of the model is compared with the actual working state of the mechanical power arm, the error is calculated and the cause is analyzed. If the error is large, the system terminal adjusts the network structure or parameters to optimize the model performance. Once the twin network model is built and verified, the system terminal exports it for subsequent simulation. By building this twin network model, various states and operating conditions of the mechanical power arm in the actual working environment can be simulated. This enables the system terminal to have a deeper understanding of the performance characteristics of the mechanical power arm and to regulate and optimize it more accurately. At the same time, the twin network model also provides the system terminal with the possibility of rehearsing path planning and grasping strategies in a virtual environment, thereby reducing the risk and cost of actual work. In general, the twin network model of the robotic arm is constructed to better simulate and optimize the performance of the robotic arm and improve its efficiency and accuracy in actual work. This step is crucial to achieve adaptive control and optimization of the robotic arm.

[0025] In the gripping preparation stage, based on the twin network model, parameter configuration is performed through the real-time operation parameters, and after completing the parameter configuration, a path planning preview simulation is performed to obtain a simulation result;

[0026] In one embodiment, in the grasping preparation stage, the system terminal uses the built twin network model for simulation. This model can simulate the movement of the mechanical power arm in the actual working environment. The system terminal configures the model parameters according to the real-time operating parameters to ensure that the state of the model is consistent with the actual mechanical power arm. After completing the parameter configuration, the twin network model performs path planning preview simulation based on its powerful learning ability. This step is very critical because it can predict and simulate various situations when the mechanical power arm moves along the planned path without actually operating the mechanical power arm. Through simulation, detailed data on the motion trajectory, speed, posture, etc. of the mechanical power arm can be obtained, and these data are the simulation results. These results can help evaluate the rationality of path planning, discover potential problems, and make corresponding optimization adjustments. In summary, using the twin network model for path planning preview simulation can effectively improve the movement efficiency and safety of the mechanical power arm and reduce the risks in actual operation.

[0027] Based on the simulation results, the motion trajectory of the mechanical power-assisting arm is optimized, and a sequence of candidate motion trajectories is obtained in order of the grasping time consumption from small to large;

[0028] In one embodiment, the system terminal conducts an in-depth analysis of the motion trajectory of the mechanical power-assisting arm based on the simulation results obtained by the twin network model. These simulation results provide the system terminal with data on the motion performance of the mechanical power-assisting arm under different paths, including key indicators such as grasping time consumption and motion stability. In the generated paths, although the simulation of the mechanical power-assisting arm has been completed, there are still problems such as inaccurate position of the grasping point, excessive or insufficient grasping force, and too fast or too slow grasping speed. In order to solve these problems, the system terminal optimizes the motion trajectory of the mechanical power-assisting arm. Specifically, the system terminal determines the target to be optimized by analyzing the simulation results and application scenario requirements. Subsequently, according to this optimization target, the motion trajectory tolerance optimization is performed, and according to the tolerance optimization results, the optimization search of the motion trajectory of the mechanical power-assisting arm is completed to obtain multiple alternative motion trajectories. Afterwards, the system terminal sorts the obtained alternative motion trajectories in ascending order according to the time consumption to form an alternative motion trajectory sequence. These trajectories can minimize the motion time and energy consumption of the mechanical power-assisting arm while ensuring the success rate of grasping. In summary, through the application of simulation results, the motion trajectory of the robotic arm was successfully optimized, and a series of efficient and reliable alternative trajectory sequences were obtained, which provided strong support for path selection and grasping strategy formulation in practical applications.

[0029] Further, based on the simulation results, the motion trajectory of the mechanical power arm is optimized, and a sequence of alternative motion trajectories is obtained in the order of small to large grasping time consumption, which also includes:

[0030] Based on the simulation results, the motion trajectory analysis is performed in combination with the application scenario requirements, and the optimization goals are set. The application scenario requirements include a first type of application scenario requirements corresponding to the automated production line and a second type of application scenario requirements corresponding to warehousing and handling;

[0031] Perform motion trajectory tolerance optimization based on the optimization target and configure a fitness function;

[0032] Preferably, based on the simulation results, the system terminal analyzes the motion trajectory performance of the mechanical power arm in different application scenarios. These application scenarios include automated production lines and warehousing and handling, and each scenario has its own specific requirements. In automated production lines, since the handling tasks are relatively fixed, the motion trajectories of the mechanical power arm are highly similar, mainly from one position to another. However, due to the strict requirements of the production line on efficiency and timing, the system terminal will pay special attention to the timing optimization of the trajectory to ensure that the mechanical power arm can complete the task accurately, thereby improving the overall production efficiency. In warehousing and handling, due to the presence of obstacles and uncertain handling requirements, the motion trajectory of the mechanical power arm is less similar. In this case, the system terminal pays more attention to the flexibility and adaptability of the trajectory. By real-time monitoring of abnormal obstacles during the handling process, the mechanical power arm can better cope with various emergencies and obstacles to ensure the smooth completion of the handling task. Subsequently, based on the above analysis, the system terminal sets the corresponding optimization target. After that, the tolerance is defined, that is, the degree to which the motion trajectory of the mechanical power arm is allowed to deviate from the ideal trajectory. The setting of the tolerance is a comprehensive consideration of factors such as the complexity of the application scenario, the accuracy requirements of the mechanical power arm, and the uncertainty in the actual working environment. Then, the system terminal starts to configure the fitness function based on the needs of the application scenario and the defined tolerance. The fitness function is used to evaluate whether the motion trajectory of the mechanical power arm meets the expectations and requirements of the system terminal. Specifically, the system terminal first calculates the deviation between the actual trajectory and the ideal trajectory to ensure that the mechanical power arm can move according to the preset path. For tasks that need to be carried strictly according to the time node, the system terminal adds the evaluation item of timing stability to ensure that the mechanical power arm can complete the operation on time. For example, a series of time windows are set to observe the completion of the operation of the mechanical power arm in each time window and judge the timing stability of the mechanical power arm. Subsequently, considering the ability of the mechanical power arm to avoid obstacles, its obstacle avoidance performance is evaluated by calculating the minimum distance between the mechanical power arm and the obstacle. Then, motion parameters such as trajectory acceleration and speed are introduced to evaluate whether the motion process of the mechanical power arm is smooth and whether there is a sudden change. After that, according to the actual application scenario requirements, the above four steps are assigned corresponding weights, and then weighted summation is performed to combine the above four steps to obtain the required fitness function. The weights in this fitness function are adjusted according to the specific needs of the application scenario. For example, in an automated production line, since the handling tasks are relatively fixed and the motion trajectories of the mechanical assistance arms are highly similar, the system terminal will correspondingly increase the weights of trajectory accuracy and timing stability, reduce the weights of obstacle avoidance and motion smoothness, or directly set the weights of obstacle avoidance and motion smoothness to 0. In summary, by configuring the fitness function, the system terminal evaluates the pros and cons of different trajectories and gradually adjusts and optimizes them according to the optimization goal.In this way, the mechanical power arm can be configured with a motion trajectory that better meets actual application requirements, achieving higher work efficiency and better performance.

[0033] Based on the fitness function, iteratively optimizing and searching the motion trajectory of the mechanical power-assisting arm to obtain a plurality of candidate motion trajectories;

[0034] Based on the multiple candidate motion trajectories, a sequence of candidate motion trajectories is obtained in order of the grasping time consumption from small to large.

[0035] Preferably, based on the configured fitness function, the system terminal performs an iterative optimization search for the motion trajectory of the mechanical power-assisting arm. During this process, the system terminal sets initialization parameters, generates an initial population based on these initialization parameters, and continuously adjusts and optimizes these initial populations to make them more in line with the requirements of the constraint information. Through multiple iterations, multiple alternative motion trajectories are obtained, and these trajectories all show good performance under the fitness function evaluation. Subsequently, these alternative motion trajectories are sorted according to the grasping time extracted when generating multiple alternative motion trajectories. The grasping time refers to the total time required for the mechanical power-assisting arm to grasp the target from the starting position and return. By comparing the grasping time of each trajectory, a sequence of alternative motion trajectories is obtained in order from small to large. The motion trajectories in this sequence are arranged according to time efficiency, which provides a basis for the subsequent selection of the best trajectory.

[0036] Furthermore, the application scenario requirements include a type of application scenario requirements corresponding to an automated production line, and also include:

[0037] According to the product characteristics on the automated production line, a first working data adjustment index is obtained, wherein the first working data adjustment index includes a gripping force and a clamping method corresponding to the product characteristics;

[0038] detecting the operating status of the automated production line in real time, and adjusting the timing of the first working data adjustment indicator in accordance with the operating status of the automated production line;

[0039] After the timing alignment adjustment, the first working data adjustment index is updated and a type of application scenario requirement is determined.

[0040] Optionally, the system terminal determines the first working data adjustment index according to the product characteristics on the automated production line. These indicators mainly include the gripping force and clamping method that match the product characteristics, which are essential for the mechanical power arm to complete the task accurately and efficiently. Subsequently, the system terminal uses the sensors and monitoring equipment on the production line to collect the operation status data of the production line in real time. These data include production speed, position of the conveyor belt, current position of the mechanical power arm, etc. After that, the production line status data collected in real time is compared with the first working data adjustment index obtained. The difference between the two is analyzed to determine whether the working data index needs to be adjusted to match the current state of the production line. If there is a difference, the system terminal makes necessary adjustments to the first working data adjustment index based on the results of the comparative analysis. The adjustment may involve gripping speed, motion trajectory, etc., to reduce or increase waiting time, ensure that the adjusted working data index is highly consistent with the operating rhythm and timing of the production line, and avoid deviations or conflicts in timing. After the timing alignment adjustment is completed, the system terminal updates the first working data adjustment index. This ensures that the motion trajectory and operation mode of the mechanical power arm always match the current production line state, thereby improving overall work efficiency. Finally, based on the updated work data adjustment indicators, the system terminal determined a category of application scenario requirements. These requirements reflect how the mechanical power arm should adjust its operation mode and parameters in specific scenarios to better adapt to the changes and requirements of the production line.

[0041] Furthermore, the application scenario requirements include two types of application scenario requirements corresponding to warehousing and transportation, and also include:

[0042] According to the warehouse layout and the distribution of goods, a second working data adjustment indicator is obtained, wherein the second working data adjustment indicator includes a gripping force and a clamping method corresponding to the type of goods;

[0043] Monitor abnormal obstacles in the handling process in real time, and adjust the trajectory of the mechanical assisting arm according to the abnormal obstacles in the handling process;

[0044] After the trajectory alignment adjustment, the second working data adjustment index is updated, and the second type of application scenario requirements are determined.

[0045] Optionally, the system terminal obtains the second work data adjustment index according to the layout of the warehouse and the distribution of the goods. These indicators are mainly determined according to different types of goods, including the force required for grabbing and the appropriate clamping method. This is done to ensure that the mechanical power arm can accurately and efficiently carry the goods according to the characteristics of different goods. During the handling process, the system terminal monitors the area that the mechanical power arm needs to pass through in real time, and detects the items in the area according to the current state information of the mechanical power arm. If the mechanical power arm is hindered from moving forward during the handling process, the item will be identified as an abnormal obstacle. Once the abnormal obstacle is determined, the system terminal will immediately align the trajectory of the motion trajectory of the mechanical power arm. This means that the movement path of the mechanical power arm will be adjusted according to the position and size of the obstacle to ensure that it can safely and smoothly bypass the obstacle and continue to complete the handling task. After completing the trajectory alignment adjustment, the system terminal updates the second work data adjustment index according to the actual handling effect and the distribution of the obstacle. In this way, the mechanical power arm can better adapt to changes in the warehouse environment and improve the accuracy and efficiency of handling. Finally, based on the updated work data adjustment index, the system terminal determines the requirements of the second application scenario. These requirements reflect how the robotic arm should adjust its operating mode and parameters in specific scenarios to meet the needs of handling different goods, providing a basis for optimization in practical applications.

[0046] Further, based on the fitness function, an iterative optimization search is performed on the motion trajectory of the mechanical power-assisting arm to obtain a plurality of candidate motion trajectories, and further comprising:

[0047] Setting initialization parameters, wherein the initialization parameters include a motion trajectory starting point and a motion trajectory end point;

[0048] Define constraint conditions, and add trajectory length and energy consumption as constraint information to the fitness function;

[0049] Based on the fitness function, combined with the initialization parameters and the constraint conditions, an initial population is generated and an iterative optimization search is performed to obtain a plurality of candidate motion trajectories.

[0050] Optionally, when optimizing the motion trajectory of the mechanical power-assisting arm, the system terminal first sets some initialization parameters, which include the starting point and the end point of the mechanical power-assisting arm. These starting points and end points are determined based on the warehouse layout, cargo distribution, and specific requirements of the handling task. Subsequently, some constraints are defined, which are used to limit and optimize the motion trajectory of the mechanical power-assisting arm. For example, if the length of the trajectory is required to be as short as possible, the movement time and energy consumption of the mechanical power-assisting arm can be reduced. Therefore, the system terminal uses the trajectory length and energy consumption as important constraint information and adds them to the fitness function. With the initialization parameters and constraints, the system terminal generates an initial motion trajectory of the mechanical power-assisting arm according to the set initialization parameters. This initial trajectory is randomly generated. Then, according to the preset population size, the above steps are repeated to generate multiple initial motion trajectories to form an initial population. After that, for each motion trajectory in the initial population, the fitness function is used to comprehensively consider the constraints such as trajectory length and energy consumption, and the corresponding weights are assigned according to the scene type for evaluation. The evaluation results reflect the performance of each trajectory under the constraints. Then, according to the results of fitness evaluation, a part of the motion trajectories with better performance are selected as parents through roulette to participate in the subsequent iterative optimization. Further, the system terminal performs a crossover operation on the selected parent trajectories, that is, combines the characteristics of two or more parent trajectories to generate new child trajectories. This helps to integrate the advantages of different trajectories. At the same time, the child trajectory is mutated, that is, a certain randomness is introduced into the child trajectory to increase the diversity of the population and avoid premature convergence to the local optimal solution. Subsequently, the new trajectory obtained after the crossover and mutation operations is added to the population to form a new generation of population. The fitness of each trajectory in the new generation of population is re-evaluated. Repeat the operations of selection, crossover, mutation and evaluation for multiple rounds of iterations until the preset number of iterations is met. After multiple rounds of iterative optimization, multiple motion trajectories with higher fitness are selected from the final population as candidate trajectories. These candidate trajectories have good performance while meeting the constraints and can be used as a reference for the actual movement of the mechanical power arm.

[0051] Based on the simulation results, the grasping strategy of the mechanical power arm is optimized, and a sequence of alternative grasping strategies is obtained in descending order of grasping success rate;

[0052] In one embodiment, the system terminal optimizes the grasping strategy of the mechanical power-assisting arm according to the results of the simulation. The grasping strategy is obtained after the simulation results are generated, combining the structural characteristics, motion principles and cargo characteristics of the mechanical power-assisting arm, covering key parameters such as different grasping forces, clamping methods and grasping paths, aiming to adapt to different grasping requirements. The strategy optimization process mainly analyzes the performance of the mechanical power-assisting arm grasping objects in the simulation to find out the shortcomings and potential improvement points. For example, if it is found that some grasping forces or clamping methods are not effective in certain situations during the simulation, resulting in a high grasping failure rate. In response to these problems, the system terminal adjusts and optimizes the grasping strategy accordingly to improve the grasping success rate. After the optimization is completed, the system terminal sorts all the alternative grasping strategies in order from large to small grasping success rates to form an alternative grasping strategy sequence. This sequence can help the system terminal quickly select the appropriate grasping strategy in actual applications to adapt to different cargo types, layouts and handling requirements.

[0053] Based on the candidate motion trajectory sequence and the candidate grasping strategy sequence, a machine learning algorithm is used to determine the optimal motion trajectory-grasping strategy combination and perform regulation and optimization through the real-time operation parameters.

[0054] In one embodiment, the system terminal combines the alternative motion trajectory sequence and the alternative grasping strategy sequence, and uses the machine learning algorithm to construct an adaptive control model to further determine the optimal combination of motion trajectory and grasping strategy. In this process, the adaptive control model takes into account the real-time operating parameters of the mechanical power arm, which reflect the actual state and environmental conditions of the mechanical power arm during operation. Through the learning and analysis of these parameters by the adaptive control model, the performance of different trajectory and strategy combinations in practical applications can be predicted, and the optimal combination can be found. Once the optimal combination is determined, the system terminal can optimize the control of the mechanical power arm to ensure that the mechanical power arm can perform the grasping task in the optimal way during actual work. This process not only improves the working efficiency of the mechanical power arm, but also reduces the error rate during operation, making the entire handling process more stable and reliable.

[0055] Further, based on the candidate motion trajectory sequence and the candidate grasping strategy sequence, a machine learning algorithm is used to determine the optimal motion trajectory-grasping strategy combination and perform control optimization through the real-time operation parameters, and further includes:

[0056] Based on the candidate motion trajectory sequence and the candidate grasping strategy sequence, construct a training data set;

[0057] Extracting features related to the motion trajectory and the grasping strategy from the real-time operation parameters to obtain a set of associated feature parameters;

[0058] Based on the training data set and the associated characteristic parameter set, a machine learning algorithm is used to construct an adaptive control model.

[0059] Optionally, a training data set is constructed based on the sequence of alternative motion trajectories and the sequence of alternative grasping strategies. This data set contains a variety of possible combinations of trajectories and strategies, providing rich samples for subsequent machine learning. Subsequently, features closely related to the motion trajectory and grasping strategy are extracted from the real-time operation parameters to form a set of associated feature parameters. These feature parameters reflect the real-time state and environmental information of the robotic arm during operation, which is crucial for determining the optimal combination of trajectory and strategy. With the training data set and the associated feature parameter set, the system terminal initializes the model parameters and sets hyperparameters such as the learning rate and the number of iterations according to the nature of the problem and the characteristics of the data. After that, the training data set is divided into a feature matrix and a target vector, wherein the feature matrix is ​​composed of an associated feature parameter set, and the target vector corresponds to the optimal combination of motion trajectory and grasping strategy. Then, through iterative optimization, the model gradually learns the mapping relationship from features to targets. During the training process, the performance indicators of the model on the training set, such as accuracy and recall, are continuously calculated to evaluate the model's fitting ability. Cross-validation is then used to evaluate the performance of the model to avoid overfitting or underfitting. Finally, the model parameters and hyperparameters are adjusted according to the evaluation results, and the model is optimized to improve the generalization ability and prediction accuracy of the model. When the number of iterations reaches the preset number of iterations, the model optimization is stopped, and the current model is output to obtain an adaptive control model. This model can automatically select the appropriate motion trajectory and grasping strategy according to the real-time operating parameters to achieve the best working effect. Through continuous learning and adjustment, the model can gradually adapt to various complex scenarios and improve the adaptability and work efficiency of the mechanical power arm.

[0060] In summary, the adaptive control method of a mechanical power-assisting arm based on multi-feature monitoring provided by the present application has the following technical effects:

[0061] The embodiment of the present application first obtains information about the mechanical power-assisting arm in the gripping preparation stage, including homing self-check, parameter setting, and target object identification and positioning. Subsequently, the mechanical power-assisting arm is subjected to feature synchronization monitoring, real-time operating parameters are collected, and handling work data records are obtained. Afterwards, a twin network model of the mechanical power-assisting arm is constructed for subsequent simulation and optimization. In the gripping preparation stage, parameter configuration is performed based on the twin network model and real-time operating parameters, and path planning preview simulation is performed. According to the simulation results, the motion trajectory of the mechanical power-assisting arm is optimized, and a sequence of alternative motion trajectories is obtained in the order of small to large grasping time consumption. At the same time, the grasping strategy is optimized to obtain a sequence of alternative grasping strategies. In the process of motion trajectory optimization, the motion trajectory is analyzed in combination with the application scenario requirements, and the optimization target is set. Then, an iterative optimization search is performed based on the fitness function to obtain multiple alternative motion trajectories. Then, the sequence of alternative motion trajectories is formed according to the grasping time consumption. After determining the sequence of alternative motion trajectories and the sequence of alternative grasping strategies, a machine learning algorithm is used to determine the optimal motion trajectory-grasping strategy combination through real-time operating parameters, and regulation optimization is performed. Targeted optimization measures are also provided for different application scenarios such as automated production lines and warehouse handling. According to product characteristics and warehouse layout, work data adjustment indicators are obtained, operating status and abnormal obstacles are detected in real time, and timing alignment and trajectory alignment adjustments are made to meet the needs of different application scenarios. These technical effects jointly solve the technical problems that the operation process of the mechanical power arm relies on fixed programming or simple feedback control, which is difficult to adapt to complex and changing working environments and task requirements, resulting in low grasping success rate and stability. It realizes comprehensive monitoring of the mechanical power arm in the grasping preparation stage, uses the twin network model to perform path planning rehearsal simulation, improves grasping accuracy, and dynamically adjusts the control strategy according to real-time operating parameters, quickly adapts to different working conditions and environmental changes, and improves the operation success rate and grasping stability.

[0062] Embodiment 2

[0063] Based on the same inventive concept as the method for adaptively controlling a mechanical power-assisting arm based on multi-feature monitoring in the aforementioned embodiment, the present application also provides an adaptively controlling device for a mechanical power-assisting arm based on multi-feature monitoring, see the attached Figure 2 , the device comprises:

[0064] Preparation phase acquisition module 1: the preparation phase acquisition module 1 is used to acquire the gripping preparation phase of the mechanical power-assisting arm, and the gripping preparation phase includes the mechanical arm homing self-check phase, the parameter setting phase, and the target object recognition and positioning phase;

[0065] Parameter acquisition module 2: the parameter acquisition module 2 is used to perform feature synchronization monitoring on the mechanical power-assisting arm, collect real-time operating parameters of the mechanical power-assisting arm, connect to the data storage unit of the mechanical power-assisting arm and obtain handling work data records, the handling work data records include position data, motion trajectory data, speed data, acceleration data, and torque data;

[0066] Model building module 3: the model building module 3 is used to build a twin network model of the mechanical power-assisting arm by comparing the mechanical power-assisting arm and recording the handling work data. The mechanical power-assisting arm includes a base component group, a bracket component group, a joint component group, a mechanical arm body, and an end-actuator component group, wherein the joint component group includes bearings, gears, and a transmission mechanism;

[0067] Simulation module 4: the simulation module 4 is used to perform parameter configuration based on the twin network model and the real-time operation parameters during the gripping preparation stage, and perform path planning preview simulation after completing the parameter configuration to obtain simulation results;

[0068] Trajectory optimization module 5: the trajectory optimization module 5 is used to optimize the motion trajectory of the mechanical power-assisting arm based on the simulation results, and obtain a sequence of candidate motion trajectories in the order of small to large grasping time consumption;

[0069] Strategy optimization module 6: the strategy optimization module 6 is used to optimize the grasping strategy of the mechanical power-assisting arm based on the simulation results, and obtain a sequence of candidate grasping strategies in descending order of grasping success rate;

[0070] The control optimization module 7 is used to determine the optimal motion trajectory-grasping strategy combination and perform control optimization based on the alternative motion trajectory sequence and the alternative grasping strategy sequence using a machine learning algorithm and through the real-time operating parameters.

[0071] Further, the trajectory optimization module 5 in the device is also used for:

[0072] Based on the simulation results, the motion trajectory analysis is performed in combination with the application scenario requirements, and the optimization goals are set. The application scenario requirements include a first type of application scenario requirements corresponding to the automated production line and a second type of application scenario requirements corresponding to warehousing and handling;

[0073] Perform motion trajectory tolerance optimization based on the optimization target and configure a fitness function;

[0074] Based on the fitness function, iteratively optimizing and searching the motion trajectory of the mechanical power-assisting arm to obtain a plurality of candidate motion trajectories;

[0075] Based on the multiple candidate motion trajectories, a sequence of candidate motion trajectories is obtained in order of the grasping time consumption from small to large.

[0076] Further, the trajectory optimization module 5 in the device is also used for:

[0077] Setting initialization parameters, wherein the initialization parameters include a motion trajectory starting point and a motion trajectory end point;

[0078] Define constraint conditions, and add trajectory length and energy consumption as constraint information to the fitness function;

[0079] Based on the fitness function, combined with the initialization parameters and the constraint conditions, an initial population is generated and an iterative optimization search is performed to obtain a plurality of candidate motion trajectories.

[0080] Further, the trajectory optimization module 5 in the device is also used for:

[0081] Based on the candidate motion trajectory sequence and the candidate grasping strategy sequence, construct a training data set;

[0082] Extracting features related to the motion trajectory and the grasping strategy from the real-time operation parameters to obtain a set of associated feature parameters;

[0083] Based on the training data set and the associated characteristic parameter set, a machine learning algorithm is used to construct an adaptive control model.

[0084] Further, the trajectory optimization module 5 in the device is also used for:

[0085] According to the product characteristics on the automated production line, a first working data adjustment index is obtained, wherein the first working data adjustment index includes a gripping force and a clamping method corresponding to the product characteristics;

[0086] detecting the operating status of the automated production line in real time, and adjusting the timing of the first working data adjustment indicator in accordance with the operating status of the automated production line;

[0087] After the timing alignment adjustment, the first working data adjustment index is updated and a type of application scenario requirement is determined.

[0088] Further, the trajectory optimization module 5 in the device is also used for:

[0089] According to the warehouse layout and the distribution of goods, a second working data adjustment indicator is obtained, wherein the second working data adjustment indicator includes a gripping force and a clamping method corresponding to the type of goods;

[0090] Monitor abnormal obstacles in the handling process in real time, and adjust the trajectory of the mechanical assisting arm according to the abnormal obstacles in the handling process;

[0091] After the trajectory alignment adjustment, the second working data adjustment index is updated, and the second type of application scenario requirements are determined.

[0092] Further, the trajectory optimization module 5 in the device is also used for:

[0093] In warehousing and handling application scenarios, synchronous monitoring is carried out to collect real-time image information;

[0094] Obtaining load information and current speed, and performing obstacle avoidance detection and early warning based on the real-time image information and comparing the motion trajectory, load information and current speed, and issuing an emergency obstacle avoidance signal;

[0095] Based on the emergency obstacle avoidance signal, an abnormal obstacle in the transportation process is determined.

[0096] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The method for adaptive control of a mechanical power-assisting arm based on multi-feature monitoring and the specific examples in Example 1 are also applicable to the adaptive control device for a mechanical power-assisting arm based on multi-feature monitoring in this example. Through the above detailed description of the method for adaptive control of a mechanical power-assisting arm based on multi-feature monitoring, those skilled in the art can clearly understand the adaptive control device for a mechanical power-assisting arm based on multi-feature monitoring in this example. Therefore, for the sake of brevity of the specification, it will not be described in detail here. For the device disclosed in the example, since it corresponds to the method disclosed in the example, the description is relatively simple, and the relevant parts can be referred to the method description.

[0097] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0098] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. Adaptive control method of mechanical power arm based on multi-feature monitoring, It is characterized in that The method comprises: Acquire the gripping preparation stage of the mechanical power-assisting arm, wherein the gripping preparation stage includes a mechanical arm homing self-checking stage, a parameter setting stage, and a target object identification and positioning stage; Performing feature synchronous monitoring on the mechanical power-assisting arm, collecting real-time operating parameters of the mechanical power-assisting arm, connecting to a data storage unit of the mechanical power-assisting arm and acquiring handling work data records, wherein the handling work data records include position data, motion trajectory data, speed data, acceleration data, and torque data; By comparing the mechanical power-assisting arm, a twin network model of the mechanical power-assisting arm is constructed through the handling work data record, wherein the mechanical power-assisting arm includes a base component group, a bracket component group, a joint component group, a mechanical arm body, and an end-actuator component group, wherein the joint component group includes bearings, gears, and a transmission mechanism; In the gripping preparation stage, based on the twin network model, parameter configuration is performed through the real-time operation parameters, and after completing the parameter configuration, a path planning preview simulation is performed to obtain a simulation result; Based on the simulation results, the motion trajectory of the mechanical power-assisting arm is optimized, and a sequence of candidate motion trajectories is obtained in order of the grasping time consumption from small to large; Based on the simulation results, the grasping strategy of the mechanical power arm is optimized, and a sequence of alternative grasping strategies is obtained in descending order of grasping success rate; Based on the candidate motion trajectory sequence and the candidate grasping strategy sequence, a machine learning algorithm is used to determine the optimal motion trajectory-grasping strategy combination and perform regulation and optimization through the real-time operation parameters; Wherein, based on the simulation results, the motion trajectory of the mechanical power arm is optimized, and a sequence of alternative motion trajectories is obtained in the order of small to large grasping time consumption, including: Based on the simulation results, the motion trajectory analysis is performed in combination with the application scenario requirements, and the optimization goals are set. The application scenario requirements include a first type of application scenario requirements corresponding to the automated production line and a second type of application scenario requirements corresponding to warehousing and handling; Perform motion trajectory tolerance optimization based on the optimization target and configure a fitness function; Based on the fitness function, iteratively optimizing and searching the motion trajectory of the mechanical power-assisting arm to obtain a plurality of candidate motion trajectories; Based on the multiple candidate motion trajectories, a sequence of candidate motion trajectories is obtained in ascending order of grasping time consumption; The application scenario requirements include a type of application scenario requirements corresponding to automated production lines, including: According to the product characteristics on the automated production line, a first working data adjustment index is obtained, wherein the first working data adjustment index includes a gripping force and a clamping method corresponding to the product characteristics; detecting the operating status of the automated production line in real time, and adjusting the timing of the first working data adjustment indicator according to the operating status of the automated production line; After the timing alignment adjustment, the first working data adjustment index is updated and a type of application scenario requirement is determined.

2. The method for adaptively controlling a mechanical power-assisting arm based on multi-feature monitoring as claimed in claim 1, It is characterized in that Based on the fitness function, an iterative optimization search is performed on the motion trajectory of the mechanical power-assisting arm to obtain a plurality of candidate motion trajectories, the method comprising: Setting initialization parameters, wherein the initialization parameters include a motion trajectory starting point and a motion trajectory end point; Define constraint conditions, and add trajectory length and energy consumption as constraint information to the fitness function; Based on the fitness function, combined with the initialization parameters and the constraint conditions, an initial population is generated and an iterative optimization search is performed to obtain a plurality of candidate motion trajectories.

3. The method for adaptively controlling a mechanical power-assisting arm based on multi-feature monitoring as claimed in claim 1, It is characterized in that Based on the candidate motion trajectory sequence and the candidate grasping strategy sequence, a machine learning algorithm is used to determine the optimal motion trajectory-grasping strategy combination and perform regulation and optimization through the real-time operation parameters, and the method includes: Based on the candidate motion trajectory sequence and the candidate grasping strategy sequence, construct a training data set; Extracting features related to the motion trajectory and the grasping strategy from the real-time operation parameters to obtain a set of associated feature parameters; Based on the training data set and the associated characteristic parameter set, a machine learning algorithm is used to construct an adaptive control model.

4. The method for adaptively controlling a mechanical power-assisting arm based on multi-feature monitoring according to claim 1, It is characterized in that The application scenario requirements include two types of application scenario requirements corresponding to warehousing and transportation, and the method further includes: According to the warehouse layout and the distribution of goods, a second working data adjustment indicator is obtained, wherein the second working data adjustment indicator includes a gripping force and a clamping method corresponding to the type of goods; Monitor abnormal obstacles in the handling process in real time, and adjust the trajectory of the mechanical assist arm according to the abnormal obstacles in the handling process; After the trajectory alignment adjustment, the second working data adjustment index is updated, and the second type of application scenario requirements are determined.

5. The method for adaptively controlling a mechanical power-assisting arm based on multi-feature monitoring as claimed in claim 4, It is characterized in that Real-time monitoring of abnormal obstacles during the handling process, and alignment and adjustment of the motion trajectory of the mechanical assisting arm according to the abnormal obstacles during the handling process, the method comprising: In warehousing and handling application scenarios, synchronous monitoring is carried out to collect real-time image information; Obtaining load information and current speed, and performing obstacle avoidance detection and early warning based on the real-time image information and comparing the motion trajectory, load information and current speed, and issuing an emergency obstacle avoidance signal; Based on the emergency obstacle avoidance signal, an abnormal obstacle in the transportation process is determined.

6. Adaptive control equipment of mechanical power arm based on multi-feature monitoring, It is characterized in that The device is used to implement the steps of the method for adaptively controlling a mechanical power-assisting arm based on multi-feature monitoring as described in any one of claims 1 to 5, the device comprising: Preparation phase acquisition module: acquires the gripping preparation phase of the robotic power-assisting arm, which includes the robotic arm homing self-check phase, parameter setting phase, and target object identification and positioning phase; Parameter acquisition module: performs feature synchronous monitoring on the mechanical power-assisting arm, collects real-time operating parameters of the mechanical power-assisting arm, connects to the data storage unit of the mechanical power-assisting arm and obtains handling work data records, wherein the handling work data records include position data, motion trajectory data, speed data, acceleration data, and torque data; Model building module: in reference to the mechanical power-assisting arm, through the handling work data record, a twin network model of the mechanical power-assisting arm is built, wherein the mechanical power-assisting arm includes a base component group, a bracket component group, a joint component group, a mechanical arm body, and an end-actuator component group, wherein the joint component group includes bearings, gears, and a transmission mechanism; Simulation module: in the gripping preparation stage, based on the twin network model, parameter configuration is performed through the real-time operation parameters, and after completing the parameter configuration, path planning preview simulation is performed to obtain simulation results; Trajectory optimization module: based on the simulation results, the motion trajectory of the mechanical power arm is optimized, and a sequence of candidate motion trajectories is obtained in the order of small to large grasping time consumption; Strategy optimization module: based on the simulation results, the grasping strategy of the mechanical power arm is optimized, and a sequence of candidate grasping strategies is obtained in descending order of grasping success rate; Control optimization module: Based on the candidate motion trajectory sequence and the candidate grasping strategy sequence, a machine learning algorithm is used to determine the optimal motion trajectory-grasping strategy combination and perform control optimization through the real-time operation parameters.

Images