A robot control method, system, electronic device and storage medium

Abstract

The application discloses a kind of robot control method, system, electronic equipment and storage medium.Therein, the method includes: according to the environmental state data and system state data of robot, the task demand of robot is determined;Based on task demand, using imitative learning optimization algorithm and quantum computing optimization algorithm, task allocation and path planning are carried out to robot, and target task allocation result and target path planning result are obtained;Based on target task allocation result and target path planning result, robot is controlled, and biological heuristic control mechanism is used to carry out cross-domain collaborative optimization to robot in the control process of robot.The technical scheme of the application improves the environmental recognition ability and three-dimensional space control ability of robot, can adapt to environmental changes faster, reduces the task execution delay caused by environmental uncertainty, improves the efficiency and quality of robot task execution.

Description

Technical Field

[0001] This invention relates to the field of robotics, and more particularly to a robot control method, system, electronic device, and storage medium. Background Technology

[0002] Robotics technology, as one of the core technologies for the development of intelligent manufacturing and service robots in the future, has developed rapidly in recent years, but some technical shortcomings have also emerged.

[0003] In existing robot control methods, when robots face newer scenarios such as complex three-dimensional spatial scenes, their recognition and control capabilities are insufficient, and they may even become lost, which seriously affects the efficiency and quality of robot task execution. Summary of the Invention

[0004] This invention provides a robot control method, system, electronic device, and storage medium to solve the problem of poor task execution efficiency and quality caused by insufficient robot recognition and control capabilities in the prior art.

[0005] According to one aspect of the present invention, a robot control method is provided, the method comprising:

[0006] Based on the robot's environmental and system status data, determine the robot's task requirements;

[0007] Based on task requirements, imitation learning optimization algorithm and quantum computing optimization algorithm are used to perform task allocation and path planning for the robot, and the target task allocation result and target path planning result are obtained.

[0008] The robot is controlled based on the target task allocation results and target path planning results, and a bio-inspired control mechanism is used to perform cross-domain collaborative optimization of the robot during the control process.

[0009] According to another aspect of the present invention, a robot control system is provided, the system comprising:

[0010] The task requirement determination module is used to determine the robot's task requirements based on the robot's environmental state data and system state data.

[0011] The task allocation and path planning module is used to allocate tasks and plan paths for the robot based on task requirements, using imitation learning optimization algorithms and quantum computing optimization algorithms, to obtain the target task allocation results and the target path planning results.

[0012] The control and cross-domain collaborative optimization module is used to control the robot based on the target task allocation results and target path planning results, and to perform cross-domain collaborative optimization of the robot by adopting a bio-inspired control mechanism during the robot control process.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] At least one processor; and

[0015] A memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the robot control method according to any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the robot control method according to any embodiment of the present invention.

[0018] The technical solution of this invention determines the robot's task requirements based on the robot's environmental and system state data. Based on these requirements, it employs imitation learning and quantum computing optimization algorithms to allocate tasks and plan paths for the robot, obtaining target task allocation and target path planning results. The robot is then controlled based on these results, and a bio-inspired control mechanism is used for cross-domain collaborative optimization during the control process. This invention, by analyzing and processing the robot's environmental and system state data, and then using imitation learning and quantum computing optimization algorithms for task allocation and path planning, solves the problems of difficulty in robot recognition and insufficient control in complex environments in existing technologies. It effectively improves the robot's environmental recognition and three-dimensional spatial control capabilities, enabling it to adapt to environmental changes more quickly and reducing task execution delays caused by environmental uncertainty. Through bio-inspired control mechanisms and cross-domain collaborative optimization, it provides the robot with more refined and natural motion control capabilities, effectively improving the efficiency and quality of robot task execution.

[0019] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of a robot control method provided according to Embodiment 1 of the present invention;

[0022] Figure 2 This is a flowchart of a robot control method provided according to Embodiment 2 of the present invention;

[0023] Figure 3 This is a schematic diagram of a robot control system according to Embodiment 3 of the present invention;

[0024] Figure 4 This is a schematic diagram of the structure of an electronic device that implements the robot control method of this invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] Example 1

[0028] Figure 1The flowchart illustrates a robot control method provided in Embodiment 1 of the present invention. This embodiment is applicable to situations requiring high-precision robot control. The method can be executed by a robot control system, which can be implemented in hardware and / or software and can be configured in an electronic device. Figure 1 As shown in the figure, the robot control method provided in this embodiment includes the following steps:

[0029] S110. Determine the robot's task requirements based on the robot's environmental and system status data.

[0030] The robot may include at least a collaborative robot; for example, in this embodiment, the robot may be a dual-arm distribution network robot. Environmental state data refers to the state data of the robot's operating environment, such as, but not limited to, obstacle locations, task locations, and work area ranges. System state data refers to the state data of the robot's own system, such as, but not limited to, the robot's position, speed, acceleration, and direction. Task requirements can be understood as information related to the tasks the robot needs to perform in the current environment, such as, but not limited to, task type, target object attributes, and operation sequence.

[0031] In this embodiment of the invention, the robot can be pre-configured with an advanced vision system, which can then control environmental perception sensors such as cameras, radar sensors, and infrared sensors configured on the robot to collect environmental state data and system state data of the robot's current environment. Then, by parsing and analyzing the aforementioned environmental state data and system state data, the current environmental characteristics and the robot's internal state are identified. Based on these characteristics and internal state, the robot's task requirements are accurately understood. These task requirements may include, but are not limited to, task type, target object attributes, and operation sequences. In one specific embodiment, the acquired environmental state data and system state data of the robot can be input into a pre-configured task recognition model. After processing by the model, the robot's task requirements are obtained.

[0032] S120. Based on task requirements, the robot is assigned tasks and planned paths using imitation learning optimization algorithm and quantum computing optimization algorithm to obtain the target task assignment results and target path planning results.

[0033] Imitation learning optimization algorithm is a method that improves a robot's decision-making ability in various complex environments by learning and imitating the behavior of experts. Its main idea is to train the robot's policy model by minimizing the difference between the robot's behavior and the expert's behavior, learn by observing the expert's decision-making process, and continuously optimize its own strategy in practical applications.

[0034] Quantum computing optimization algorithms are algorithms that utilize the principles of quantum computing to solve optimization problems. They accelerate the solution process by leveraging properties of quantum mechanics, such as quantum superposition and quantum entanglement. Quantum computing optimization algorithms can include, but are not limited to: quantum simulated annealing, quantum random walk, variational quantum algorithms, quantum dynamic optimization, quantum incremental optimization, and quantum machine learning algorithms. These algorithms enable global search in non-convex optimization problems, thereby finding the global optimum.

[0035] In this embodiment of the invention, an imitation learning optimization algorithm can be invoked to control the robot to imitate and learn from the demonstration data of human experts. By minimizing the error of the robot when imitating human actions, the imitation accuracy and environmental adaptability of the robot can be improved. Then, based on the current task requirements, a quantum computing optimization algorithm is invoked to solve the quantum optimization problem, obtaining the frequency distribution of the optimal solution. Subsequently, the robot is assigned tasks and its path is planned according to the frequency distribution of the optimal solution, resulting in the corresponding target task allocation results and target path planning results. The quantum optimization problem is to minimize the overall energy consumption and time of the robot in the process of performing the task, while ensuring the accuracy of the actions and environmental adaptability.

[0036] S130. The robot is controlled based on the target task allocation results and target path planning results, and a bio-inspired control mechanism is used to perform cross-domain collaborative optimization of the robot during the control process.

[0037] Bioinspired control mechanisms can refer to algorithms that design robot control and planning by simulating evolutionary processes or behavioral patterns in the biological world, in order to achieve more efficient intelligent robot control. Bioinspired control mechanisms may include, but are not limited to, artificial neural networks, genetic algorithms, and particle swarm optimization algorithms.

[0038] In this embodiment of the invention, the robot can be controlled using the aforementioned determined target task allocation results and target path planning results. At the same time, during the robot's task execution, a pre-configured bio-inspired control mechanism is used to perform cross-domain collaborative optimization of the robot, providing the robot with more refined and natural motion control capabilities, which greatly improves the accuracy and safety of task execution.

[0039] Furthermore, during the robot's task execution, the operating status of each component can be acquired to determine if any malfunctions have occurred. Based on the malfunction status of each component, dynamic strategy adjustments and task reallocation can be implemented. Specifically, if all robot components are operating normally, performance monitoring is continuously performed, and user or operator feedback is collected to adjust workflows and task configurations promptly. If some robot components experience minor malfunctions that do not significantly affect the robot's overall functionality, the system quickly identifies and locates the malfunctioning component, performs a preliminary diagnosis to determine the cause and nature of the malfunction, and reduces the workload of affected components based on the severity of the malfunction, reallocating critical tasks to other normally operating components or robots. If some robot components experience serious malfunctions, the affected components or the entire robot operation is immediately stopped, provided safety is ensured. A thorough fault diagnosis is rapidly conducted to determine the cause, nature, and urgency of the malfunction, and a corresponding repair or replacement plan is developed. Based on the function and urgency of the malfunctioning component, the current work tasks are reorganized, and backup systems or solutions are activated. After completing the necessary repairs, comprehensive testing is conducted to ensure the malfunction is effectively controlled.

[0040] The technical solution of this invention determines the robot's task requirements based on the robot's environmental and system state data. Based on these requirements, it employs imitation learning and quantum computing optimization algorithms to allocate tasks and plan paths for the robot, obtaining target task allocation and target path planning results. The robot is then controlled based on these results, and a bio-inspired control mechanism is used for cross-domain collaborative optimization during the control process. This invention, by analyzing and processing the robot's environmental and system state data, and then using imitation learning and quantum computing optimization algorithms for task allocation and path planning, solves the problems of difficulty in robot recognition and insufficient control in complex environments in existing technologies. It effectively improves the robot's environmental recognition and three-dimensional spatial control capabilities, enabling it to adapt to environmental changes more quickly and reducing task execution delays caused by environmental uncertainty. Through bio-inspired control mechanisms and cross-domain collaborative optimization, it provides the robot with more refined and natural motion control capabilities, effectively improving the efficiency and quality of robot task execution.

[0041] Example 2

[0042] Figure 2 This is a flowchart of a robot control method provided in Embodiment 2 of the present invention. It is further optimized and extended based on the above embodiments and can be combined with various optional technical solutions in the above embodiments. For example... Figure 2As shown in the figure, the robot control method provided in this embodiment includes the following steps:

[0043] S210. Control the pre-configured environmental perception sensors to collect environmental state data and system state data of the robot's current environment; wherein, the environmental perception sensors include at least one of the following: camera, infrared sensor, lidar, light sensor, microphone and vibration sensor, the environmental state data includes at least: obstacle position and task location, and the system state data includes at least: position, speed and acceleration.

[0044] Among them, environmental perception sensors can refer to sensor devices used to collect environmental state data and system state data of robots. It can be understood that corresponding environmental perception sensors can be configured for robots according to the actual task requirements. For example, environmental perception sensors include at least one of the following: camera, infrared sensor, lidar, light sensor, microphone and vibration sensor.

[0045] In this embodiment of the invention, during the robot's operation, various pre-configured environmental perception sensors can be controlled to collect environmental state data and system state data of the robot's current environment, so that the above data can be analyzed to determine the robot's task requirements in the current environment; wherein, the environmental perception sensors include at least one of the following: camera, infrared sensor, lidar, light sensor, microphone and vibration sensor.

[0046] S220. Call the pre-configured task recognition model to determine the robot's task requirements for environmental state data and system state data; wherein, the task requirements include at least: task type, target object attributes and operation sequence.

[0047] The task recognition model can be understood as a pre-trained network model used to determine the corresponding task requirements based on the robot's environmental state data and system state data. The task recognition model can include at least a self-learning model, such as a self-supervised learning model and a Transformer model.

[0048] In this embodiment of the invention, after acquiring the robot's environmental state data and system state data, the data can be input into a pre-trained task recognition model. This model analyzes and understands the robot's environmental and system state data to identify the robot's current task requirements. The task requirements include at least the following information: task type, target object attributes, and operation sequence. The task type can refer to the type of task the robot is to perform, such as live wiring, live wire disconnection, or grounding ring installation. Target object attributes can include information such as the location, shape, and mass of obstacles. The operation sequence refers to the relevant information about the execution steps of the task. For example, in a live wiring task, the operation sequence could include: identifying the target wire to be worked on; determining the wiring location; and controlling the robotic arm to strip the wire insulation and complete the fastening.

[0049] Furthermore, based on the above embodiments of the invention, it also includes:

[0050] In response to user input commands, the robot parses the task requirements based on environmental and system status data.

[0051] Specifically, it can also receive robot control instructions, i.e., user input instructions, manually input by users or operators based on the robot's current state. By parsing these user input instructions, the robot's task requirements in response to environmental and system state data can be obtained.

[0052] S230. Obtain the pre-created expert dataset, call the imitation learning optimization algorithm, and use the demonstration data in the expert dataset to perform imitation learning on the robot.

[0053] In this context, an expert dataset can be understood as a dataset containing a series of demonstration data from experts (demonstrators). The demonstration data may include at least state and action information; by analyzing this expert-generated state and action information, the robot can learn how to perform a specific task.

[0054] In this embodiment of the invention, a pre-created expert dataset containing demonstration data can be obtained, the operational action information of the expert demonstration in the demonstration data can be captured, and then the imitation learning optimization algorithm can be called to control the robot to perform action analysis and imitation learning.

[0055] Furthermore, based on the above embodiments of the invention, the objective in the imitation learning process is to minimize the error of the robot when imitating human actions, and the corresponding objective function is configured as follows:

[0056]

[0057] Where T represents the length of the observation time window, that is, the time period for evaluating the robot's imitation performance; M represents the set of environmental state vectors perceived by the robot. E λ represents the number of environmental states; λ represents the time decay factor, indicating that the importance of early actions gradually decreases over time; N represents the number of action dimensions, i.e., the position and orientation of each joint in space; f n (x t x represents a function that represents the robot's action n at time t; t This represents the robot's system state vector at time t, including position, velocity, and acceleration. α represents the actual value of the human-demonstrated action n at time t; α(S,R) represents the environmental adaptability weight function, used to measure the robot's adaptability to environmental changes; S represents the set of ideal response state vectors corresponding to the set of environmental state vectors; i σ represents the i-th environmental state perceived by the robot. C The standard deviation, representing the difference in environmental conditions, is used to control the sensitivity of the environmental adaptability component. (C) IL The range of (T,S) is [0, +∞), when C IL When (T,S)=0, it means the robot perfectly imitates human actions and is completely adapted to the environment; when C IL When (T,S) increases, it indicates that the accuracy of the imitation decreases or the adaptability to the environment deteriorates.

[0058] The function f of the robot performing action n n (x t It can be defined as follows:

[0059] f n (x t )=β·g(x t ,p n )+γ·d(x t ,a n )

[0060] In the formula, p n represents the parameters of action n under ideal conditions, i.e., target joint angles or target position parameters; β and γ represent weight parameters, respectively representing the importance of the main part of the action execution and the adjustment part of the action; g(x t ,p n ) represents the robot's relationship with the target parameter p when performing a specific action n. n The distance measurement function for the difference between them; d(x t ,a n ) represents the difference function between the robot's current state and the ideal adaptive state for action n; a nThis represents the fitness parameter for action n.

[0061] The environmental adaptability weight function α(S,R) can be defined as follows:

[0062]

[0063] In the formula, ω i σ represents the weight associated with the i-th environmental state, indicating the importance of the current state to the overall task; c,i s represents the standard deviation corresponding to the i-th environmental state, used to control the sensitivity of the weights to the influence of environmental changes; i -r i This represents the Euclidean distance between the actual environmental state and the ideal response state; when the actual environmental state s i Approaching the ideal response state r indefinitely i At that time, s i -r i A value approaching 1 indicates that the robot is highly adapted to its environment; conversely, if the value deviates beyond the preset range, then s i -r i A rapid decrease in the value indicates a decline in the robot's adaptability to the environment. This can be addressed by adjusting the weight ω for each environmental state. i and sensitivity σ c,i Adjust the robot's environmental adaptability to specific application scenarios.

[0064] S240. The quantum optimization problem, optimization parameters and task requirements corresponding to the quantum computing optimization algorithm are quantum encoded, and the quantum computing optimization algorithm is called to iteratively solve the task allocation result and path planning result corresponding to the robot to obtain the target task allocation result and target path planning result.

[0065] In this embodiment of the invention, leveraging the superior parallel computing capabilities, quantum parallel search, randomness, and ability to handle nonlinear problems of quantum computing optimization algorithms, the task allocation and path planning problems of robots can be solved using quantum computing optimization algorithms. Specifically, a quantum optimization problem is first defined, with the objective of minimizing the overall energy consumption and time of the robot during task execution, while ensuring action accuracy and environmental adaptability. The corresponding objective optimization function is as follows:

[0066]

[0067] In the formula, Λ represents the control coefficient of environmental adaptability; p represents the parameter of action n under ideal conditions.

[0068] When it is necessary to optimize the task allocation and path planning of robots, the above quantum optimization problem can be formalized and corresponding optimization parameters can be defined, such as robot position, obstacle position, work area range, task location and time limit parameters. Then, the above optimization parameters and task requirements are converted into qubit form for encoding, ensuring that the possible states and values ​​of each parameter can be represented in quantum computing.

[0069] When quantum algorithms are chosen to solve problems, quantum circuits are constructed based on the characteristics of the optimization problem, including the selection and parameter configuration of quantum gates. If the quantum circuits are successfully implemented and run, the frequency distribution of the optimal solution is obtained through repeated measurements, and the robot's action strategy is formulated or adjusted based on the results, thereby obtaining the corresponding target task allocation results and target path planning results. Subsequently, the robot can be dynamically controlled according to the target task allocation results and target path planning results.

[0070] S250: Control the robot to execute the corresponding task and planned path according to the target task allocation result and the target path planning result.

[0071] In this embodiment of the invention, after the robot's task allocation and path planning strategies are determined by the imitation learning optimization algorithm and the quantum computing optimization algorithm, that is, after obtaining the target task allocation result and the target path planning result, the robot can be controlled according to the target task allocation result and the target path planning result.

[0072] Furthermore, when the robot performs a task according to the strategy obtained by the quantum computing optimization algorithm, the interaction between the robot and the environment can be continuously monitored and feedback data during the execution process can be collected. If more than one of the following situations is found: environmental changes, changes in task requirements, or robot performance not meeting expectations, data can be collected again, the parameters of the optimization problem can be updated, and the quantum computing optimization algorithm can be called again for iterative optimization until the optimal performance standard is achieved.

[0073] S260. During the robot's task execution, a bio-inspired control model based on artificial neural networks is invoked to adjust and control the robot's muscles and joints.

[0074] Artificial Neural Networks (ANNs) are mathematical models that use a structure similar to the synaptic connections of the brain to process information. These models have the following characteristics: high parallelism, associative memory, and good fault tolerance.

[0075] In this embodiment of the invention, a bio-inspired control model based on artificial neural networks can be used to optimize the robot's muscles and joints during the robot's task execution, thereby making the robot's movements smooth and natural without any stuttering.

[0076] Furthermore, based on the above embodiments of the invention, the bio-inspired control model can be expressed as:

[0077]

[0078] in, A set of signal vectors representing the input to an artificial neural network, used to mimic the input signals of the human nervous system; This represents the set of weight vectors corresponding to each input signal of an artificial neural network, used to simulate the intensity of information transmission between neurons;

[0079] This represents the set of parameters corresponding to the robot's muscle and joint control, used to mimic the forces of muscle contraction and joint movement; and These represent the robot's current muscle and joint states, as well as the desired target state; σ H represents the normalization function, used to standardize the output of artificial neural networks; tanh represents the hyperbolic tangent function, used to simulate the activation function of human neurons; b a δ represents the bias corresponding to the a-th neuron, used to adjust the neuron's activation threshold; δ represents the sensitivity of muscle and joint responses; Ω represents the operating domain or time range of the entire system. The Gaussian function representing the response of simulated biological muscles and joints is used to reflect the difference between the actual state and the target state.

[0080] S270. Call the pre-configured cross-domain collaborative optimization model to optimize and adjust the application field, task module and tools corresponding to the robot.

[0081] In this context, "application area" can refer to different robot task execution environments. "Task module" can refer to the tasks that the robot can perform, such as live wiring, live wire disconnection, and grounding ring installation. "Tools" can refer to the tools and equipment required for the robot to perform tasks, such as wire strippers, wrenches, wire cutters, and nut breakers.

[0082] In this embodiment of the invention, during the robot's task execution, a pre-configured cross-domain collaborative optimization model can be used to optimize and adjust the robot's corresponding application domain, task modules, and tools, thereby optimizing the robot's adaptability to the environment and task execution efficiency.

[0083] Furthermore, based on the above embodiments of the invention, the cross-domain collaborative optimization model can be expressed as:

[0084]

[0085] Where D represents the number of different application areas; Q = {Q1, Q2, ..., Q} H} represents a set of different task modules, Q h Represents the h-th task module; G = {G1, G2, ..., G...} H} represents the set of tools corresponding to the set of task modules Q, and G represents the set of tools. h Indicates the h-th tool; Ω q Φ represents the parameter space for task execution in the q-th application domain; q (θ q ,g q (Q,G)) represents the performance evaluation function corresponding to the q-th application domain; θq represents the domain-specific parameters; g q (Q,G) represents the task-tool matching function, expressing the efficiency and adaptability of task and tool configuration in domain q; λ1 represents the regularization parameter, used to balance the importance of task execution efficiency and task-tool matching degree; H represents the total number of task modules; Ψ(Q h G h ) represents the penalty function.

[0086] Among them, the performance evaluation function Φ q (θ q ,g q (Q,G) is defined as follows:

[0087]

[0088] In the formula, α q and β q θ represents the weighting coefficients, which adjust for the impact of domain-specific parameter matching and task tool configuration adaptability on performance evaluation; q θ represents the actual parameters for task execution within the current domain. q,target Indicates the target parameters for task execution in the current domain; g q,target This represents the ideal configuration of tasks and tools; σ θ and σ g These are the sensitivity to control parameter matching and configuration adaptability, respectively.

[0089] Penalty function Ψ(Q) h G h The definition is as follows:

[0090]

[0091] In the formula, δ k σ represents the weight of the mismatch between the k-th task and the tool, and σ represents the importance of different types of mismatch in the overall penalty. QG A parameter indicating the sensitivity to mismatch between control tasks and tools.

[0092] Furthermore, based on the above embodiments of the invention, it also includes:

[0093] The robot's strategy is dynamically adjusted by calling the pre-configured fault detection function and task reallocation function respectively.

[0094] In this embodiment of the invention, in order to dynamically adjust the robot's strategy, the following fault detection function F(L,J) can be introduced:

[0095]

[0096] In the formula, L={l1,l2,...,l N} represents the set of current states of all parts of the robot; J = {j1,j2,...,j...} N} represents the set of normal operating state parameters for each component; σ f The parameter represents the sensitivity of fault detection. The fault detection function F(L,J) determines the sum of deviations between the robot's current state and its normal state. When the current state of a component deviates significantly from its normal state, the value of the corresponding term will approach 1, indicating that the current component may have malfunctioned.

[0097] After determining the fault detection function, a dynamic adjustment strategy and task reallocation function R(T,F) can be further defined:

[0098]

[0099] In the formula, T = {t1, t2, ..., t} M} represents the original task allocation; F = {f1, f2, ..., f M} represents the fault state of the component corresponding to the task, which can be determined by the fault detection function F. c This indicates that the value is 0 when the component is functioning normally and close to 1 when the component malfunctions.

[0100] During the robot's task execution, the working status of each component can be acquired to determine if any malfunctions have occurred. Based on the malfunction status of each component, the robot's dynamic adjustment strategy and task reallocation can then be implemented. Specifically, this can be divided into the following three categories:

[0101] ①When all f cWhen all values ​​are 0, it means that all parts of the robot are within the normal operating range and no significant signs of failure are found. In this case, continue to implement routine system monitoring and maintenance procedures, including regular inspections, performance monitoring and preventive maintenance, continuously record the robot's operating data and system status for long-term performance analysis and optimization, regularly evaluate the efficiency and quality of the robot in completing tasks, and collect feedback from users or operators in order to adjust workflows and task configurations in a timely manner.

[0102] ② When one or more f c When the value is less than the critical value but not 0, it indicates that some parts of the robot have begun to have minor faults, but have not yet seriously affected the overall function of the robot. The system quickly identifies and locates the parts with minor faults, performs a preliminary diagnosis to determine the cause and nature of the fault, and reduces the workload of the affected parts according to the degree of the fault. The critical tasks are redistributed to other normally functioning parts or robots. At the same time, the identified parts with minor faults are temporarily repaired or adjusted, a long-term maintenance or replacement plan is planned, and the monitoring of the repaired parts is strengthened and the corresponding data is recorded.

[0103] ③ When one or more f c When the value exceeds the critical value, it indicates that some components of the robot have suffered serious malfunctions. Under the premise of ensuring safety, immediately stop the affected components or stop the operation of the entire robot. At the same time, quickly conduct in-depth fault diagnosis to determine the cause, nature and urgency of the fault, formulate corresponding repair or replacement plans, reorganize the current work tasks according to the function and urgency of the faulty components, and activate the backup system or backup plan. After completing the necessary repair work, conduct comprehensive testing to ensure that the fault has been effectively controlled.

[0104] Next, comparative tests will be conducted to demonstrate the advantages of the technical solution of this invention in terms of adaptability, operational efficiency, recognition capability, and safety. Two robots were selected for comparison: one is a dual-arm robot applying the technical solution of this invention, and the other is a dual-arm robot employing traditional path planning and environmental perception technologies.

[0105] The test environment was designed as a three-dimensional space containing various obstacles to simulate a complex industrial production line environment. Multiple task points were set up within the experimental environment, requiring the robot to start from the starting point, complete operations at all task points, and finally return to the starting point. Operations at each task point included picking up, placing, or manipulating a specific object.

[0106] The comparative test consisted of multiple rounds. In each round, both robots performed the same task sequence, and the performance of each robot was evaluated. The experimental results are shown in the table below:

[0107]

[0108] As shown in the table above, the dual-arm robot using the technical solution of this invention exhibits superior performance compared to traditional dual-arm robots, demonstrating its high adaptability and control capabilities in complex three-dimensional spaces, which is particularly evident when facing varying lighting conditions and obstacles. By integrating an advanced vision system and multi-source environmental perception sensors, it can capture minute changes in the environment in real time and quickly adjust task allocation and path planning strategies through imitation learning optimization algorithms and quantum computing optimization algorithms. Furthermore, the application of bio-inspired control mechanisms and cross-domain collaborative optimization provides the robot with more refined and natural motion control capabilities, greatly improving the accuracy and safety of task execution.

[0109] The technical solution of this invention collects environmental and system state data of the robot's environment through multi-source environmental perception sensors, enabling real-time capture of minute environmental changes. Through imitation learning optimization and quantum computing optimization algorithms, it rapidly adjusts task allocation and path planning strategies, improving the robot's environmental recognition and three-dimensional spatial control capabilities. This allows the robot to adapt to environmental changes more quickly, reducing task execution delays caused by environmental uncertainty and decreasing reliance on human intervention. The robot can complete tasks independently with less human input, demonstrating a high level of autonomy and intelligence. Furthermore, through bio-inspired control mechanisms and cross-domain collaborative optimization, the robot is provided with more refined and natural motion control capabilities, effectively improving the efficiency and quality of task execution. Simultaneously, it effectively avoids potential safety accidents, ensuring the safety of the human-robot coexisting environment.

[0110] Example 3

[0111] Figure 3 This is a schematic diagram of a robot control system provided in Embodiment 3 of the present invention. Figure 3 As shown, the system includes:

[0112] The task requirement determination module 31 is used to determine the task requirements of the robot based on the robot's environmental state data and system state data.

[0113] The task allocation and path planning module 32 is used to perform task allocation and path planning on the robot based on the task requirements, using imitation learning optimization algorithm and quantum computing optimization algorithm, to obtain target task allocation results and target path planning results.

[0114] The control and cross-domain collaborative optimization module 33 is used to control the robot based on the target task allocation result and the target path planning result, and to perform cross-domain collaborative optimization of the robot using a bio-inspired control mechanism during the control process.

[0115] The technical solution of this invention involves a task requirement determination module that determines the robot's task requirements based on the robot's environmental and system state data. A task allocation and path planning module, based on these requirements, uses imitation learning and quantum computing optimization algorithms to allocate tasks and plan paths for the robot, obtaining target task allocation and target path planning results. A control and cross-domain collaborative optimization module controls the robot based on these results, employing a bio-inspired control mechanism for cross-domain collaborative optimization during the control process. This invention, by analyzing and processing the robot's environmental and system state data, and then using imitation learning and quantum computing optimization algorithms for task allocation and path planning, solves the problems of difficulty in robot recognition and insufficient control in complex environments in existing technologies. It effectively improves the robot's environmental recognition and three-dimensional spatial control capabilities, enabling it to adapt to environmental changes more quickly and reducing task execution delays caused by environmental uncertainty. Through bio-inspired control mechanisms and cross-domain collaborative optimization, it provides the robot with more refined and natural motion control capabilities, effectively improving the efficiency and quality of robot task execution.

[0116] Furthermore, based on the above embodiments of the invention, the task requirement determination module 31 includes:

[0117] A state data acquisition unit is used to control pre-configured environmental perception sensors to acquire environmental state data and system state data of the robot's current environment; wherein, the environmental perception sensors include at least one of the following: camera, infrared sensor, lidar, light sensor, microphone and vibration sensor, the environmental state data includes at least: obstacle position and task location, and the system state data includes at least: position, velocity and acceleration;

[0118] The task requirement determination unit is used to call a pre-configured task recognition model to determine the task requirements of the robot in response to the environmental state data and the system state data; wherein, the task requirements include at least: task type, target object attributes, and operation sequence.

[0119] Furthermore, based on the above embodiments of the invention, the task allocation and path planning module 32 includes:

[0120] The imitation learning unit is used to acquire a pre-created expert dataset, call the imitation learning optimization algorithm, and use the demonstration data in the expert dataset to perform imitation learning on the robot.

[0121] The quantum optimization unit is used to quantum encode the quantum optimization problem, optimization parameters and task requirements corresponding to the quantum computing optimization algorithm, and call the quantum computing optimization algorithm to iteratively solve the task allocation result and path planning result corresponding to the robot to obtain the target task allocation result and the target path planning result.

[0122] Furthermore, based on the above embodiments of the invention, the objective function configured during the imitation learning process, which aims to minimize the error of the robot in imitating human actions, is as follows:

[0123]

[0124] Where T represents the length of the observation time window; M represents the set of environmental state vectors perceived by the robot. E λ represents the number of environmental states; N represents the time decay factor; f represents the number of action dimensions; n (x t x represents a function that represents the robot's action n at time t; t This represents the system state vector of the robot at time t; α represents the actual value of the human-demonstrated action n at time t; α(S,R) represents the environmental adaptation weighting function. S represents the set of ideal response state vectors corresponding to the set of environmental state vectors; i σ represents the i-th environmental state perceived by the robot. C Standard deviation representing differences in environmental conditions;

[0125] The quantum optimization problem is to minimize the robot's total energy consumption and time during task execution, while ensuring action accuracy and environmental adaptability. The corresponding objective optimization function is:

[0126]

[0127] Where Λ represents the control coefficient for environmental adaptability; p represents the parameter of action n under ideal conditions.

[0128] Furthermore, based on the above embodiments of the invention, the control and cross-domain collaborative optimization module 33 includes:

[0129] The first control unit is used to control the robot to execute the corresponding task and the planned path according to the target task allocation result and the target path planning result;

[0130] The second control unit is used to invoke a bio-inspired control model based on artificial neural networks to adjust and control the muscles and joints of the robot during the task execution process.

[0131] The cross-domain collaborative optimization unit is used to call a pre-configured cross-domain collaborative optimization model to optimize and adjust the application domain, task module and tools corresponding to the robot.

[0132] Furthermore, based on the above embodiments of the invention, the bio-inspired control model is expressed as follows:

[0133]

[0134] in, This represents the set of signal vectors input to an artificial neural network.

[0135] This represents the set of weight vectors corresponding to each input signal of an artificial neural network.

[0136] This represents the set of parameters corresponding to the robot's muscle and joint control.

[0137] and These represent the robot's current muscle and joint states, as well as the desired target state; σ H tanh represents the normalization function; b represents the hyperbolic tangent function. a δ represents the bias corresponding to the a-th neuron; δ represents the sensitivity of muscle and joint response; Ω represents the operating domain or time range of the entire system; A Gaussian function representing the simulated response of biological muscles and joints;

[0138] The cross-domain collaborative optimization model is expressed as follows:

[0139]

[0140] Where D represents the number of different application areas; Q = {Q1, Q2, ..., Q} H} represents a set of different task modules, Q h Represents the h-th task module; G = {G1, G2, ..., G...} H} represents the set of tools corresponding to the set of task modules Q, and G represents the set of tools. h Indicates the h-th tool; Ω q Φ represents the parameter space for task execution in the q-th application domain; q (θ q ,g q (Q,G)) represents the performance evaluation function corresponding to the q-th application domain; θq represents the domain-specific parameters; g q (Q,G) represents the task-tool matching function; λ1 represents the regularization parameter; H represents the total number of task modules; Ψ(Q hG h ) represents the penalty function.

[0141] Furthermore, based on the above embodiments, the control and cross-domain collaborative optimization module 33 further includes:

[0142] The dynamic strategy adjustment unit is used to call the pre-configured fault detection function and task reallocation function respectively to dynamically adjust the strategy of the robot.

[0143] The robot control system provided in the embodiments of the present invention can execute the robot control method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0144] Example 4

[0145] Figure 4 A schematic diagram of the structure of an electronic device 40 that can be used to implement embodiments of the present invention is shown. For example... Figure 4 As shown, the electronic device 40 includes at least one processor 41 and a memory, such as a read-only memory (ROM) 42 or a random access memory (RAM) 43, communicatively connected to the at least one processor 41. The memory stores computer programs executable by the at least one processor. The processor 41 can perform various appropriate actions and processes based on the computer program stored in the ROM 42 or loaded from storage unit 48 into the RAM 43. The RAM 43 can also store various programs and data required for the operation of the electronic device 40. The processor 41, ROM 42, and RAM 43 are interconnected via a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.

[0146] Multiple components in electronic device 40 are connected to I / O interface 45, including: input unit 46, such as keyboard, mouse, etc.; output unit 47, such as various types of monitors, speakers, etc.; storage unit 48, such as disk, optical disk, etc.; and communication unit 49, such as network card, modem, wireless transceiver, etc. Communication unit 49 allows electronic device 40 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0147] Processor 41 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 41 performs the various methods and processes described above, such as robot control methods.

[0148] In some embodiments, the robot control method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or mounted on electronic device 40 via ROM 42 and / or communication unit 49. When the computer program is loaded into RAM 43 and executed by processor 41, one or more steps of the robot control method described above may be performed. Alternatively, in other embodiments, processor 41 may be configured to execute the robot control method by any other suitable means (e.g., by means of firmware).

[0149] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0150] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0151] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0152] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0153] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0154] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0155] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0156] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A robot control method, characterized in that, The method includes: The task requirements of the robot are determined based on the robot's environmental and system status data. Based on the task requirements, the robot is assigned tasks and planned paths using imitation learning optimization algorithm and quantum computing optimization algorithm to obtain the target task assignment results and target path planning results. The robot is controlled based on the target task allocation result and the target path planning result, and a bio-inspired control mechanism is used to perform cross-domain collaborative optimization of the robot during the control process. The step of controlling the robot based on the target task allocation result and the target path planning result, and employing a bio-inspired control mechanism to perform cross-domain collaborative optimization of the robot during the control process, includes: The robot is controlled to execute the corresponding task and the planned path according to the target task allocation result and the target path planning result; During the robot's task execution, a bio-inspired control model based on artificial neural networks is invoked to adjust and control the robot's muscles and joints; The pre-configured cross-domain collaborative optimization model is invoked to optimize and adjust the application domain, task modules, and tools corresponding to the robot; The bio-inspired control model is represented as follows: ； in, This represents the set of signal vectors input to an artificial neural network. This represents the set of weight vectors corresponding to each input signal of an artificial neural network. This represents the set of parameters corresponding to the robot's muscle and joint control. and These represent the robot's current muscle and joint states, as well as the desired target state; tanh represents the normalization function; b represents the hyperbolic tangent function. a This represents the bias corresponding to the a-th neuron; Indicates the sensitivity of muscle and joint responses; Indicates the operating domain or time range of the entire system; A Gaussian function representing the simulated response of biological muscles and joints; The cross-domain collaborative optimization model is expressed as follows: ； Where D represents the number of different application areas; This represents a collection of different task modules. This represents the h-th task module; This represents the set of tools corresponding to the set of task modules Q. This represents the h-th tool; This represents the parameter space for task execution in the q-th application domain; This represents the performance evaluation function corresponding to the q-th application domain; Indicates domain-specific parameters; This represents a function for matching tasks with tools; This represents the regularization parameter; H represents the total number of task modules. This represents the penalty function.

2. The method according to claim 1, characterized in that, The step of determining the robot's task requirements based on the robot's environmental state data and system state data includes: The system controls pre-configured environmental perception sensors to collect environmental state data and system state data of the robot's current environment; wherein, the environmental perception sensors include at least one of the following: camera, infrared sensor, lidar, light sensor, microphone and vibration sensor, the environmental state data includes at least: obstacle position and task location, and the system state data includes at least: position, velocity and acceleration; A pre-configured task recognition model is invoked to determine the robot's task requirements in response to the environmental state data and the system state data; wherein, the task requirements include at least: task type, target object attributes, and operation sequence.

3. The method according to claim 1, characterized in that, Based on the task requirements, the robot is assigned tasks and its path is planned using an imitation learning optimization algorithm and a quantum computing optimization algorithm, resulting in target task assignment results and target path planning results, including: Obtain a pre-created expert dataset, and invoke the imitation learning optimization algorithm and use the demonstration data in the expert dataset to perform imitation learning on the robot; The quantum optimization problem, optimization parameters, and task requirements corresponding to the quantum computing optimization algorithm are quantum encoded, and the quantum computing optimization algorithm is called to iteratively solve the task allocation results and path planning results corresponding to the robot to obtain the target task allocation results and the target path planning results.

4. The method according to claim 3, characterized in that, In the imitation learning process, the objective is to minimize the error of the robot when imitating human actions. The corresponding objective function is: ； Where T represents the length of the observation time window; M represents the set of environmental state vectors perceived by the robot. E For the number of environmental states; Indicates the time decay factor; N represents the number of action dimensions; The function representing the action n performed by the robot at time t; x t This represents the system state vector of the robot at time t; This represents the actual value of the action n demonstrated by the human at time t; This represents the environmental adaptability weighting function; S represents the set of ideal response state vectors corresponding to the set of environmental state vectors; i This represents the i-th environmental state perceived by the robot; Standard deviation representing differences in environmental conditions; The quantum optimization problem is to minimize the robot's total energy consumption and time during task execution, while ensuring action accuracy and environmental adaptability. The corresponding objective optimization function is: ； in, The control coefficient represents the environmental adaptability; p represents the parameter of action n under ideal conditions.

5. The method according to claim 1, characterized in that, Also includes: The robot's dynamic strategy is adjusted by calling the pre-configured fault detection function and task reallocation function respectively.

6. A robot control system, characterized in that, The system includes: The task requirement determination module is used to determine the task requirements of the robot based on the robot's environmental state data and system state data. The task allocation and path planning module is used to allocate tasks and plan paths for the robot based on the task requirements, using imitation learning optimization algorithm and quantum computing optimization algorithm, to obtain target task allocation results and target path planning results. The control and cross-domain collaborative optimization module is used to control the robot based on the target task allocation result and the target path planning result, and to perform cross-domain collaborative optimization of the robot using a bio-inspired control mechanism during the control process. The control and cross-domain collaborative optimization module includes: The first control unit is used to control the robot to execute the corresponding task and the planned path according to the target task allocation result and the target path planning result; The second control unit is used to invoke a bio-inspired control model based on artificial neural networks to adjust and control the muscles and joints of the robot during the task execution process. The cross-domain collaborative optimization unit is used to call a pre-configured cross-domain collaborative optimization model to optimize and adjust the application domain, task module and tools corresponding to the robot; The bio-inspired control model is represented as follows: ； in, This represents the set of signal vectors input to an artificial neural network. This represents the set of weight vectors corresponding to each input signal of an artificial neural network. This represents the set of parameters corresponding to the robot's muscle and joint control. and These represent the robot's current muscle and joint states, as well as the desired target state; tanh represents the normalization function; b represents the hyperbolic tangent function. a This represents the bias corresponding to the a-th neuron; Indicates the sensitivity of muscle and joint responses; Indicates the operating domain or time range of the entire system; A Gaussian function representing the simulated response of biological muscles and joints; The cross-domain collaborative optimization model is expressed as follows: ； Where D represents the number of different application areas; This represents a collection of different task modules. This represents the h-th task module; This represents the set of tools corresponding to the set of task modules Q. This represents the h-th tool; This represents the parameter space for task execution in the q-th application domain; This represents the performance evaluation function corresponding to the q-th application domain; Indicates domain-specific parameters; This represents a function for matching tasks with tools; This represents the regularization parameter; H represents the total number of task modules. This represents the penalty function.

7. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the robot control method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the robot control method according to any one of claims 1-5.

Images