Walking control method of humanoid robot and related equipment

Abstract

The embodiment of the invention provides a walking control method of a humanoid robot and related equipment, and belongs to the technical field of robots. The method comprises the following steps: receiving a production work order from a manufacturing execution system; acquiring real-time environment data, and according to the real-time environment data, constructing an environment model including a ground state, an obstacle, a robot state and a task target area; based on the environment model and the task nodes allocated to the current humanoid robot, performing multi-objective optimization by using a model prediction control algorithm, and generating an optimal control strategy including a walking path and a gait; the current foot type humanoid robot is controlled to execute the optimal control strategy, and the optimal control strategy is dynamically adjusted through a task execution dynamic adjustment model according to the environment model; and synchronously updating the evolved experience segments to all humanoid robots through a feedback learning model. According to the embodiment of the invention, the humanoid robot can achieve a stable, flexible and intelligent execution effect in a changeable production scene.

Description

Technical Field

[0001] This application relates to the field of robotics technology, and in particular to a walking control method and related equipment for a humanoid robot. Background Technology

[0002] Among related technologies, with the development of industrial automation and intelligent manufacturing, the application of legged humanoid robots in flexible manufacturing has received widespread attention. Legged humanoid robots possess strong terrain adaptability and flexibility, enabling them to perform tasks in uneven or dynamically changing environments. Compared to traditional wheeled or tracked robots, legged humanoid robots have significant advantages in complex terrains, capable of traversing obstacles, climbing stairs, and walking on irregular surfaces. Therefore, legged humanoid robots have been widely used for task execution in complex indoor and outdoor environments, such as automated warehousing, material handling, and assembly. Summary of the Invention

[0003] The main objective of this application is to propose a walking control method and related equipment for humanoid robots, which enables humanoid robots to achieve stable, flexible and intelligent execution in changing production scenarios.

[0004] To achieve the above objectives, one aspect of this application proposes a walking control method for a humanoid robot, the method comprising: The system receives production work orders from the manufacturing execution system, parses the production work orders to obtain multiple task nodes, and, based on the task nodes, uses a multi-robot collaboration and task allocation model to conduct task bidding and dynamic allocation among multiple humanoid robots according to the current state and task priority of each humanoid robot. The system acquires real-time environmental data and constructs an environmental model based on this data, including ground conditions, obstacles, the robot's own state, and the target area of ​​the task. Based on the environmental model and the task nodes assigned to the current humanoid robot, a model predictive control algorithm is used to perform multi-objective optimization to generate an optimal control strategy that includes walking path and gait; wherein, the multi-objective optimization considers at least task priority and environmental uncertainty; The system controls the current legged humanoid robot to execute the optimal control strategy, and dynamically adjusts the optimal control strategy according to the environment model through a task execution dynamic adjustment model. By using a feedback learning model, key data during task execution is extracted as experience fragments, encoded into genes, and stored in a population gene bank. Then, through a gene evolution model, the experience fragments in the population gene bank are screened, cross-linked, and mutated, and the evolved experience fragments are synchronously updated to all humanoid robots.

[0005] In some embodiments, the step of using the task node, through a multi-robot collaboration and task allocation model, to conduct task bidding and dynamic allocation among multiple humanoid robots based on the current state and task priority of each humanoid robot includes: Publish the multiple task nodes to the task marketplace; Each humanoid robot is controlled to generate candidate execution plans for task nodes in the task market and bid on them based on its current position, load status, tool availability, and task execution capabilities. The multi-robot collaboration and task allocation model comprehensively evaluates the bidding schemes of each legged humanoid robot, task priorities, and the principle of avoiding path conflicts, so as to dynamically allocate the task nodes to the most suitable humanoid robot.

[0006] In some embodiments, controlling each humanoid robot to generate candidate execution plans for task nodes in the task market and to bid on them based on its current position, load status, tool availability, and task execution capabilities includes: Control each humanoid robot to scan the surrounding environment in real time to obtain environmental perception data; Based on the environmental perception data and the requirements of the task nodes, one or more candidate execution schemes are evaluated and generated; the candidate execution schemes include at least the estimated path planning, gait strategy and task completion time.

[0007] In some embodiments, the step of using a model predictive control algorithm to perform multi-objective optimization based on the environment model and the task node assigned to the current humanoid robot to generate an optimal control strategy including walking path and gait includes: Construct a prediction model based on the aforementioned humanoid robot dynamics model; Define a multi-objective cost function that includes path tracking error, gait stability, energy consumption, and penalty terms related to task priority; In each control cycle, with the current humanoid robot state and the environmental model as initial conditions, the optimal control input sequence for a future period is obtained by solving the minimization problem of the multi-objective cost function within the set prediction time domain. The first control input in the optimal control input sequence is transmitted to the current humanoid robot, and the steps are repeated in the next control cycle. In each control cycle, with the current humanoid robot state and the environment model as initial conditions, the optimal control input sequence for a future period is obtained by solving the minimization problem of the multi-objective cost function within the set prediction time domain, so as to calculate the optimal control strategy including walking path and gait.

[0008] In some embodiments, the penalty term related to task priority in the multi-objective cost function is configured such that, for high-priority task nodes, the path tracking error term corresponding to the high-priority task node has a weight coefficient higher than a preset weight threshold, so as to drive the humanoid robot to complete high-priority tasks first.

[0009] In some embodiments, controlling the current legged humanoid robot to execute the optimal control strategy, and dynamically adjusting the optimal control strategy according to the environment model through a task execution dynamic adjustment model, includes: If a new obstacle or a change in the position of the task target is detected during the humanoid robot's task execution, path replanning is immediately triggered to update the environment model, resulting in an updated environment model. Based on the updated environment model, the optimal control strategy is recalculated through the task execution dynamic adjustment model. If a task priority change instruction is received from the manufacturing execution system, the order or path planning of the currently executed tasks will be dynamically adjusted according to the changed task priority.

[0010] In some embodiments, the step of screening, crossover, and mutation operations on experience fragments in the population gene pool using a gene evolution model, and synchronously updating the evolved experience fragments to all humanoid robots, includes: Based on the success rate, execution efficiency, or energy consumption indicators of historical tasks, the fitness of the empirical fragments in the population gene pool is evaluated. Based on the fitness evaluation results, experience fragments with high fitness are selected for retention; The selected experience fragments are subjected to crossover and mutation operations to generate the evolved experience fragments and update them to all humanoid robots.

[0011] To achieve the above objectives, another aspect of this application provides a walking control device for a humanoid robot, the device comprising: The task processing module is used to receive production work orders from the manufacturing execution system, parse the production work orders to obtain multiple task nodes, and based on the task nodes, through a multi-robot collaboration and task allocation model, to conduct task bidding and dynamic allocation among multiple humanoid robots according to the current status and task priority of each humanoid robot. The environment construction module is used to acquire real-time environmental data and construct an environment model that includes ground conditions, obstacles, the robot's own state, and the task target area based on the real-time environmental data. The strategy optimization module is used to perform multi-objective optimization based on the environment model and the task nodes assigned to the current humanoid robot, using a model predictive control algorithm to generate an optimal control strategy that includes walking path and gait; wherein, the multi-objective optimization considers at least task priority and environmental uncertainty; The strategy adjustment module is used to control the current legged humanoid robot to execute the optimal control strategy, and to dynamically adjust the optimal control strategy according to the environment model through the task execution dynamic adjustment model; The experience evolution module is used to extract key data from the task execution process into experience fragments through a feedback learning model, encode them into genes and store them in a population gene bank. The gene evolution model is used to screen, crossover and mutate the experience fragments in the population gene bank and synchronously update the evolved experience fragments to all humanoid robots.

[0012] To achieve the above objectives, another aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described above.

[0013] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described above.

[0014] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer program product, including a computer program that, when executed by a processor, implements the aforementioned method.

[0015] The embodiments of this application include at least the following beneficial effects: This application provides a walking control method, device, electronic device, storage medium, and program product for a humanoid robot. This solution receives production work orders from a manufacturing execution system, parses the work orders to obtain multiple task nodes, and based on these task nodes, uses a multi-robot collaboration and task allocation model to conduct task bidding and dynamic allocation among multiple humanoid robots according to their current state and task priority. It also acquires real-time environmental data and constructs an environmental model based on this data, including ground conditions, obstacles, the robot's own state, and the task target area. Based on the environmental model and the task allocation to the current humanoid robot... The robot's task nodes utilize model predictive control algorithms for multi-objective optimization, generating an optimal control strategy that includes walking paths and gait. This multi-objective optimization considers at least task priority and environmental uncertainty. The system controls the current legged humanoid robot to execute the optimal control strategy, and dynamically adjusts this strategy based on the environmental model through a task execution dynamic adjustment model. A feedback learning model extracts key data from the task execution process into experience fragments, encodes them genetically, and stores them in a population gene bank. A gene evolution model then filters, cross-references, and mutates these experience fragments in the gene bank, synchronously updating them to all humanoid robots. Implementing this embodiment, by integrating production work order parsing, multi-robot dynamic task allocation, real-time environmental modeling and model predictive control optimization, and a feedback-driven gene evolution mechanism, significantly enhances the overall performance of humanoid robots in flexible manufacturing environments. It enables seamless collaboration between task scheduling and path planning, flexibly responding to multi-task demands and workstation switching through task bidding and allocation mechanisms. This effectively overcomes the limitations of separating these two processes in traditional methods, enabling humanoid robots to achieve stable, flexible, and intelligent execution in diverse production scenarios. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of an implementation environment provided in an embodiment of this application; Figure 2 This is a flowchart of the walking control method for a humanoid robot provided in an embodiment of this application; Figure 3 This is a schematic diagram of the walking control process of a humanoid robot in one embodiment; Figure 4 This is a schematic diagram of the walking control device for a humanoid robot provided in an embodiment of this application; Figure 5 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0018] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”

[0019] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0021] Before providing a detailed description of the embodiments of this application, some of the nouns and terms involved in the embodiments of this application will be explained first. The nouns and terms involved in the embodiments of this application are subject to the following interpretations.

[0022] 1) Model predictive control (MPC) is an advanced control strategy widely used in industrial process control and the optimization of complex systems. Its core idea is to predict the future state of the system and dynamically adjust the current control input to achieve optimal control performance.

[0023] In flexible manufacturing environments, legged humanoid robots face even greater challenges. Because flexible manufacturing demands highly flexible and adaptable production systems to cope with rapidly changing production tasks, frequent workstation changes, and diverse product requirements, traditional path planning and gait control methods often focus on optimizing static or pre-defined task scenarios, making it difficult to maintain sufficient robustness and control accuracy in dynamic and changing working environments. Model predictive control, as a commonly used dynamic optimization algorithm, has been applied in areas such as walking control, path planning, and force control of legged humanoid robots. By predicting the future state of the system and implementing optimal control based on the current state and objectives, it has significant advantages in stability and accuracy; however, it still has limitations in complex and changing flexible manufacturing scenarios.

[0024] Traditional model predictive control (MMCC) typically assumes a known environmental state and lacks adaptability to dynamic environmental changes and unknown obstacles, limiting the actual flexibility of humanoid robots. Furthermore, many related technologies separate task scheduling and path planning, lacking an integrated mechanism for task scheduling and path optimization in multi-task, multi-robot collaborative work, making it difficult to meet the demands of flexible manufacturing for parallel multi-task operation, rapid workstation switching, and resource sharing. In addition, in complex terrain or multi-robot collaborative scenarios, the high computational cost of MMCC affects real-time response speed, becoming a bottleneck for efficient system operation in dynamic production environments. Therefore, optimizing existing MMCC algorithms to improve their overall performance in flexible manufacturing environments has become an important research direction in the field of legged humanoid robot control.

[0025] In view of this, this application provides a walking control method for a humanoid robot, which enables the humanoid robot to achieve stable, flexible and intelligent execution in changing production scenarios.

[0026] The walking control method for a humanoid robot provided in this application relates to the field of robotics. This walking control method for a humanoid robot can be applied to a terminal, a server, or software running on a terminal or server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or vehicle terminal, but is not limited to these. The server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network. The software can be an application implementing the walking control method for the humanoid robot, but is not limited to the above forms.

[0027] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0028] It should be noted that in all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirection to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0029] like Figure 1 The diagram shown is a schematic representation of an implementation environment provided in an embodiment of this application. (Refer to...) Figure 1 The implementation environment includes at least one terminal 102 and a server 101. The terminal 102 and the server 101 can be connected via a network, either wirelessly or via a wired connection, to complete data transmission and exchange.

[0030] Server 101 can be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0031] Additionally, server 101 can also be a node server in a blockchain network. Blockchain is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms.

[0032] Terminal 102 can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. Terminal 102 and server 101 can be directly or indirectly connected via wired or wireless communication, and this embodiment of the application does not impose any limitations.

[0033] For example, based on Figure 1 The implementation environment shown in this application embodiment provides a walking control method for a humanoid robot. The following description uses the application of this walking control method for a humanoid robot in server 101 as an example. It can be understood that the walking control method for a humanoid robot can also be applied to terminal 102.

[0034] Figure 2 This is an optional flowchart of the walking control method for a humanoid robot provided in the embodiments of this application. The execution subject of the walking control method for the humanoid robot can be any of the aforementioned electronic devices (including servers or terminals). Figure 2 The method may include, but is not limited to, steps S201 to S205.

[0035] Step S201: Receive production work orders from the manufacturing execution system, parse the production work orders to obtain multiple task nodes, and based on the task nodes, use a multi-robot collaboration and task allocation model to conduct task bidding and dynamic allocation among multiple humanoid robots according to the current status and task priority of each humanoid robot.

[0036] In some embodiments, the humanoid robot is communicatively connected to an electronic device that can communicate with a Manufacturing Execution System (MES). Upon receiving a production work order from the MES, the electronic device can parse the work order. Specifically, the production task is issued by the MES to the electronic device and then transmitted to the robot. A task is typically a production work order, which contains multiple processes and task nodes. For example, a work order might include tasks such as "material handling," "parts assembly," and "quality inspection." The electronic device can parse the production work order and break it down into multiple specific task nodes, forming a task flowchart. Each task node includes not only a task description but also information such as task execution requirements, workstation location, and tooling requirements.

[0037] After task nodes are generated, the electronic devices will sort the tasks according to factors such as priority, time requirements, and complexity, and publish each task node to the task marketplace. Each humanoid robot will assess whether it is suitable to perform a task based on its current working status (such as location, workload, tool availability, etc.), submit a task bidding proposal, and wait for task allocation.

[0038] Optionally, the electronic device can publish multiple task nodes to the task market; control each humanoid robot to generate candidate execution plans for the task nodes in the task market and bid for them based on its current position, load status, tool availability and task execution capability; through a multi-robot collaboration and task allocation model, the bidding plans of each legged humanoid robot, task priority and the principle of avoiding path conflicts are comprehensively evaluated to dynamically allocate task nodes to the most suitable humanoid robot.

[0039] In the task market, multiple humanoid robots bid on tasks based on their proposed solutions. Electronic devices evaluate the execution potential of each solution based on factors such as the robot's task execution capabilities, current location, tool availability, and task priority. Each robot then submits a bid based on its resources and workload.

[0040] The task marketplace comprehensively evaluates the task plans of various robots and selects the most suitable humanoid robot for task allocation. In scenarios involving multi-robot collaboration, the system further optimizes task allocation, making collaboration between humanoid robots more efficient and avoiding resource or path conflicts. For example, in material handling tasks, multiple robots may need to share workbenches and tools. The system dynamically adjusts task allocation based on factors such as the urgency and difficulty of the task, ensuring smooth and seamless collaboration between multiple robots.

[0041] The combination of distributed autonomous decision-making and centralized optimization and coordination fully leverages the heterogeneous states and capabilities of each humanoid robot, stimulating efficiency potential through competitive bidding. Simultaneously, considerations of task prioritization and conflict prevention ensure the priority execution of critical tasks and the orderly operation of the overall system. This achieves a shift from optimizing the efficiency of individual humanoid robots to maximizing the collaborative benefits of multiple robots, thereby enhancing the agility and overall collaborative efficiency of the entire flexible manufacturing system in response to dynamic, multi-task production scenarios.

[0042] Furthermore, each humanoid robot is controlled to scan the surrounding environment in real time to obtain environmental perception data; based on the environmental perception data and the requirements of the task nodes, one or more candidate execution schemes are evaluated and generated; the candidate execution schemes include at least the estimated path planning, gait strategy and task completion time.

[0043] Upon receiving a task node, the humanoid robot activates its perception module and scans the surrounding environment in real time. This module uses various sensors, such as vision, lidar, IMU, and force sensors, to perceive the environment and acquire data on the current terrain, feasible landing points, obstacles, and material locations. Based on this perception data, the robot generates one or more candidate execution plans.

[0044] The electronic devices combine task node requirements with perceived data to evaluate the feasibility of each execution plan. At this point, the humanoid robot not only considers path planning but also changes in the current environment. Especially when facing uneven ground or obstacles, the humanoid robot dynamically adjusts its gait and path to ensure the accuracy and stability of task execution. Finally, the humanoid robot generates the optimal execution plan and submits it to the task marketplace to await task allocation.

[0045] Step S202: Obtain real-time environmental data and construct an environmental model based on the real-time environmental data, including ground conditions, obstacles, robot's own state, and the task target area.

[0046] In some embodiments, electronic devices can continuously acquire real-time environmental data such as depth information, point cloud data, and the robot's own pose and motion state from multiple sources of sensors, including depth cameras, LiDAR, inertial measurement units, and joint encoders, configured in the humanoid robot. Based on this fused perception data, an environmental model is constructed. This environmental model is a digital spatiotemporal representation, the core of which lies in integrating multimodal information. The ground state includes terrain geometry features, slope, friction coefficient, and other drivability analysis; obstacles are represented by the position, contour, and motion prediction of dynamic and static objects in the model; the humanoid robot's own state includes real-time pose, joint angles, velocity, and dynamic constraints; and the task target area defines the spatial coordinates and posture requirements of the task destination.

[0047] Step S203: Based on the environment model and the task nodes assigned to the current humanoid robot, multi-objective optimization is performed using a model predictive control algorithm to generate an optimal control strategy that includes walking path and gait; wherein, multi-objective optimization considers at least task priority and environmental uncertainty.

[0048] In some embodiments, a prediction model based on a humanoid robot dynamics model is constructed; a multi-objective cost function is defined, including path tracking error, gait stability, energy consumption, and penalty terms related to task priority; in each control cycle, using the current humanoid robot state and environment model as initial conditions, the optimal control input sequence for a future period is obtained by solving the minimization problem of the multi-objective cost function within a set prediction time domain; the first control input in the optimal control input sequence is transmitted to the current humanoid robot, and the steps are repeated in the next control cycle. In each control cycle, using the current humanoid robot state and environment model as initial conditions, the optimal control input sequence for a future period is obtained by solving the minimization problem of the multi-objective cost function within a set prediction time domain, so as to calculate the optimal control strategy including walking path and gait.

[0049] By closely integrating prediction, multi-objective optimization, and real-time feedback, humanoid robots can proactively plan their movements in complex environments, actively balance multiple sometimes conflicting objectives such as accurate arrival, stable walking, and energy efficiency, and dynamically adjust their behavioral focus according to task priorities. This significantly improves the adaptability, robustness, and overall task execution efficiency of humanoid robots in dynamic scenarios such as flexible manufacturing, both theoretically and practically.

[0050] Furthermore, the penalty term related to task priority in the multi-objective cost function is configured such that for high-priority task nodes, the path tracking error term corresponding to the high-priority task node has a weight coefficient higher than a preset weight threshold, so as to drive the humanoid robot to complete high-priority tasks first.

[0051] Through this weighting configuration, electronic devices can autonomously make decisions that meet the production cycle requirements in complex scenarios with multiple tasks running in parallel. For example, they can prioritize rushing to emergency workstations or handling critical processes. This significantly improves the intelligence of task execution and the ability to support the overall efficiency of the manufacturing system while ensuring the stability of overall control performance.

[0052] Step S204: Control the current legged humanoid robot to execute the optimal control strategy, and dynamically adjust the optimal control strategy according to the environment model through the task execution dynamic adjustment model.

[0053] In some embodiments, if a new obstacle or a change in the position of the task target is detected during the humanoid robot's task execution, path replanning is immediately triggered to update the environment model, resulting in an updated environment model. Based on the updated environment model, the optimal control strategy is recalculated by dynamically adjusting the model through task execution. If a task priority change instruction is received from the manufacturing execution system, the order or path planning of the currently executed task is dynamically adjusted according to the changed task priority.

[0054] Once the task is assigned, the humanoid robot begins execution. At this point, the robot generates an execution path and begins moving based on the requirements of the task nodes and real-time perception data. Using the MPC control algorithm, the robot adjusts its gait, speed, and position in real time to ensure precise task execution. During execution, the robot collects environmental feedback through sensors; for example, a torque sensor monitors the force of the movements, a vision sensor identifies obstacles, and an IMU provides feedback on the robot's posture and stability.

[0055] During execution, if the humanoid robot encounters sudden obstacles or uneven ground, its electronic devices dynamically adjust the robot's current path, using the MPC algorithm to recalculate the optimal path and gait, ensuring the task is unaffected and completed on time. Through this process, the humanoid robot can flexibly respond to various environmental changes, maintaining high precision and stability while performing its tasks.

[0056] Step S205: Through the feedback learning model, key data in the task execution process is extracted into experience fragments, which are then encoded into genes and stored in the population gene bank. Through the gene evolution model, the experience fragments in the population gene bank are screened, cross-linked, and mutated, and the evolved experience fragments are synchronously updated to all humanoid robots.

[0057] In some embodiments, electronic devices can continuously monitor the task execution process of each humanoid robot, extracting key data such as operation sequences for successfully coping with specific terrain or efficiently completing a certain type of assembly, controller parameters, and environmental state correlation information, and encapsulating this data into standardized experience fragments. Electronic devices can also genetically encode these experience fragments, employing a structured expression method similar to biological gene sequences to transform complex experiences into data units that can be stored, copied, combined, and modified, and aggregated and stored in a population gene bank.

[0058] Optionally, the fitness of experience fragments in the population gene pool is evaluated based on the success rate, execution efficiency, or energy consumption indicators of historical tasks; based on the fitness evaluation results, experience fragments with high fitness are selected for retention; crossover and mutation operations are performed on the selected experience fragments to generate evolved experience fragments and update them to all humanoid robots.

[0059] After completing the task, the humanoid robot records and extracts key data from the execution process, such as path planning, execution parameters, energy consumption, and task completion quality. This data forms experience fragments, which are then encoded into genes according to a preset format. Experience fragments include information such as task type, environmental conditions, specific actions performed, and whether the task was successful. The humanoid robot stores this experience data in an experience gene bank and continuously optimizes its execution strategy based on this data.

[0060] After each task, the electronic device compares the humanoid robot's performance under different tasks, selects the best-performing execution strategies and experience fragments, and optimizes the experience base through genetic evolution algorithms. This allows the humanoid robot to continuously evolve, improving its execution efficiency and accuracy in similar tasks. This provides strong support for the long-term learning and optimization of the humanoid robot, ensuring that it can achieve higher efficiency and fewer errors in future tasks.

[0061] After humanoid robots perform tasks and accumulate experience, the population gene bank is updated regularly based on task performance feedback. The robots convert their successful experiences into gene fragments and upload them to the gene bank. The system uses genetic algorithms to screen, crossover, and mutate these experience fragments, preserving superior genes and optimizing performance.

[0062] The evolved genes will be synchronously updated in all humanoid robots, enabling each robot to quickly learn and apply collective experience, thereby improving the overall task execution efficiency of the group. In multi-robot collaborative tasks, the updated gene pool allows each robot to learn from collective experience, not only improving the execution capabilities of individual humanoid robots but also making multi-robot collaboration more efficient.

[0063] The system architecture of the electronic device mainly includes six core modules: perception and environment modeling module, task scheduling and path planning module, optimized MPC control module, multi-robot collaboration and task allocation module, feedback and learning module, and task execution and dynamic adjustment module. Each module collaborates with each other to complete the dynamic walking control and task execution of the humanoid robot in a flexible manufacturing environment.

[0064] The perception and environment modeling module is the foundational module of the system, responsible for real-time acquisition of various information about the robot's environment. This module utilizes multiple sensors (such as vision sensors, LiDAR, force sensors, and IMU sensors) to acquire environmental data and generates real-time maps through environment modeling. It employs an attention mechanism to extract crucial information from the vast amount of sensor data, particularly potential future landing sites, and uses this data to support subsequent path planning and task execution. This module can also perceive dynamic changes in the environment, such as temporary material stacking, the appearance of obstacles, or changes in the workbench, ensuring that the robot can adjust its control strategy promptly according to environmental changes.

[0065] The task scheduling and path planning module is responsible for generating task execution plans and movement paths based on production task requirements and environmental conditions. This module first decomposes tasks into multiple task nodes according to the production work order and prioritizes these nodes. In the case of multiple tasks executing in parallel, the system dynamically adjusts task allocation and path planning based on multiple factors such as task priority, execution complexity, and environmental constraints. The system employs an optimized path planning algorithm to ensure the robot avoids obstacles and chooses the shortest path possible, improving task execution efficiency. This module also adjusts path planning based on changes in the production line (such as workstation switching and tool replacement) to ensure the robot can complete tasks efficiently and accurately.

[0066] The optimized MPC control module is the core module of this invention. By combining real-time perception data, task scheduling information, and environmental models, it optimizes the robot's walking path, gait, and task execution strategy. Unlike traditional MPC algorithms, the MPC control module of this invention introduces a multi-objective optimization mechanism, considering factors such as task priority, workload, and environmental uncertainties in addition to optimizing stability and speed. The system predicts future states and calculates the optimal control input in real time based on the current robot state, target state, and constraints. The MPC module can adjust the robot's gait according to feedback, ensuring stable walking in complex environments and dynamically adjusting the path according to external disturbances.

[0067] The multi-robot collaboration and task allocation module ensures intelligent task allocation and efficient resource utilization when multiple robots work collaboratively. In flexible manufacturing environments, multiple robots typically need to collaborate to complete tasks such as material handling and parts assembly. This module dynamically adjusts task allocation based on each robot's current load, position, and available resources. By optimizing task allocation and path coordination, it ensures efficient collaboration among multiple robots, avoiding path conflicts and resource competition. Furthermore, this module supports multi-robot collaboration strategy optimization, dynamically optimizing collaboration schemes based on factors such as robot collaboration needs, resource sharing, and task priorities.

[0068] The feedback and learning module is a crucial part of the system's adaptive optimization. By collecting feedback data (such as execution time, energy consumption, and task completion rate) in real time during task execution, this module provides continuous optimization suggestions to the robot. This feedback data is transformed into experience fragments and stored in an experience gene bank, providing a reference for future tasks. Through reinforcement learning and other learning algorithms, the robot extracts successful execution experiences from historical tasks and optimizes its future behavioral strategies. After each task execution, the system adaptively optimizes the MPC algorithm based on feedback information, thereby improving the robot's efficiency and accuracy in task execution.

[0069] The task execution and dynamic adjustment module ensures that the robot can dynamically adjust itself based on real-time environmental changes and task feedback during task execution. During execution, the robot continuously updates its task execution status through sensor data and adjusts its path or gait according to environmental changes (such as changes in the workbench or the appearance of obstacles). For example, when the robot is performing a material handling task, the system may detect new obstacles on the path or changes in the target location. In this case, the task execution module will immediately adjust the robot's path to ensure efficient task completion. This module works closely with other modules (such as path planning and MPC control modules) to ensure the stability and accuracy of task execution.

[0070] Figure 3 This is a schematic diagram of the walking control process of a humanoid robot in one embodiment, such as... Figure 3 As shown, the electronic equipment first issues work orders and decomposes tasks, then performs robot perception and task generation (including sensor data acquisition and task execution plan generation), production work order parsing, and task priority ranking in parallel. Next, it enters the task allocation and multi-robot collaboration stage, covering task plan bidding and task allocation and scheduling. In the subsequent execution and dynamic adjustment phase, if no obstacles are encountered, the task continues to be executed; if obstacles are encountered, obstacle detection is performed first, followed by dynamic path adjustment or path replanning and execution, with path optimization, gait adjustment, and real-time feedback adjustments during execution. After the task is completed, it enters the experience extraction and gene encoding stage, involving task execution result recording, task execution parameter extraction, experience fragment encoding, and then updating the population gene bank, while simultaneously performing gene fragment evolution and gene update distribution.

[0071] Steps S201 to S205 as illustrated in this embodiment involve receiving a production work order from the manufacturing execution system, parsing the work order to obtain multiple task nodes, and, based on these task nodes, using a multi-robot collaboration and task allocation model, bidding for and dynamically allocating tasks among the humanoid robots according to their current state and task priority. Real-time environmental data is acquired, and an environmental model including ground conditions, obstacles, the robot's own state, and the task target area is constructed based on this data. Based on the environmental model and the task nodes allocated to the current humanoid robot, a model predictive control algorithm is used for multi-objective optimization to generate an optimal control strategy including the walking path and gait. This multi-objective optimization considers at least task priority and environmental uncertainty. The current legged humanoid robot is controlled to execute the optimal control strategy, and the model is dynamically adjusted through task execution. Based on the environmental model, the system dynamically adjusts the optimal control strategy. Through a feedback learning model, key data during task execution is extracted as experience fragments, encoded genetically, and stored in a population gene bank. A gene evolution model is then used to screen, crossover, and mutate these experience fragments, which are then synchronously updated to all humanoid robots. By integrating production work order parsing, multi-robot dynamic task allocation, real-time environmental modeling and model predictive control optimization, and a feedback-driven gene evolution mechanism, the system significantly enhances the overall performance of humanoid robots in flexible manufacturing environments. It enables seamless collaboration between task scheduling and path planning, and flexibly responds to multi-task demands and workstation switching through task bidding and allocation mechanisms. This effectively overcomes the limitations of separating these two processes in traditional methods, allowing humanoid robots to achieve stable, flexible, and intelligent execution in diverse production scenarios.

[0072] Please see Figure 4 This application also provides a walking control device for a humanoid robot, which can implement the above-described method. The device includes: The task processing module 401 is used to receive production work orders from the manufacturing execution system, parse the production work orders to obtain multiple task nodes, and based on the task nodes, through a multi-robot collaboration and task allocation model, to conduct task bidding and dynamic allocation among multiple humanoid robots according to the current status and task priority of each humanoid robot. The environment construction module 402 is used to acquire real-time environmental data and construct an environment model that includes ground conditions, obstacles, robot's own conditions, and the task target area based on the real-time environmental data. The strategy optimization module 403 is used to perform multi-objective optimization based on the environment model and the task nodes assigned to the current humanoid robot, using a model predictive control algorithm to generate an optimal control strategy that includes walking path and gait; wherein, the multi-objective optimization considers at least task priority and environmental uncertainty; The strategy adjustment module 404 is used to control the current legged humanoid robot to execute the optimal control strategy, and to dynamically adjust the optimal control strategy according to the environment model through the task execution dynamic adjustment model; The experience evolution module 405 is used to extract key data from the task execution process into experience fragments through a feedback learning model, encode them into genes and store them in the population gene bank. Through the gene evolution model, the experience fragments in the population gene bank are screened, cross-linked and mutated, and the evolved experience fragments are synchronously updated to all humanoid robots.

[0073] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0074] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0075] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0076] Please see Figure 5 , Figure 5 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 501 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 502 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 502 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 502 and is called and executed by the processor 501 using the methods described in the embodiments of this application. The input / output interface 503 is used to implement information input and output; The communication interface 504 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 505 transmits information between various components of the device (e.g., processor 501, memory 502, input / output interface 503, and communication interface 504); The processor 501, memory 502, input / output interface 503, and communication interface 504 are connected to each other within the device via bus 505.

[0077] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0078] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0079] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0080] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0081] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0082] The humanoid robot walking control method, device, electronic device, storage medium, and program product provided in this application embodiment receive production work orders from a manufacturing execution system, parse the production work orders to obtain multiple task nodes, and based on the task nodes, use a multi-robot collaboration and task allocation model to conduct task bidding and dynamic allocation among multiple humanoid robots according to the current state and task priority of each humanoid robot; acquire real-time environmental data, and construct an environmental model including ground state, obstacles, robot's own state, and task target area based on the real-time environmental data; based on the environmental model and the task nodes allocated to the current humanoid robot, use a model predictive control algorithm to perform multi-objective optimization. The system generates an optimal control strategy that includes walking paths and gait. Multi-objective optimization considers at least task priority and environmental uncertainty. It controls the current legged humanoid robot to execute the optimal control strategy and dynamically adjusts the strategy based on the environmental model through a task execution dynamic adjustment model. A feedback learning model extracts key data from the task execution process into experience fragments, encodes them genetically, and stores them in a population gene bank. A gene evolution model then filters, cross-references, and mutates these experience fragments in the population gene bank, and synchronously updates the evolved experience fragments to all humanoid robots. This enables humanoid robots to achieve stable, flexible, and intelligent execution in diverse production scenarios.

[0083] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0084] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0085] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0086] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0087] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application 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 this application 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.

[0088] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0090] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0092] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0093] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A walking control method of a humanoid robot, characterized by, The method includes the following steps: The system receives production work orders from the manufacturing execution system, parses the production work orders to obtain multiple task nodes, and, based on the task nodes, uses a multi-robot collaboration and task allocation model to conduct task bidding and dynamic allocation among multiple humanoid robots according to the current state and task priority of each humanoid robot. The system acquires real-time environmental data and constructs an environmental model based on this data, including ground conditions, obstacles, the robot's own state, and the target area of ​​the task. Based on the environmental model and the task nodes assigned to the current humanoid robot, a model predictive control algorithm is used to perform multi-objective optimization to generate an optimal control strategy that includes walking path and gait; wherein, the multi-objective optimization considers at least task priority and environmental uncertainty; The system controls the current legged humanoid robot to execute the optimal control strategy, and dynamically adjusts the optimal control strategy according to the environment model through a task execution dynamic adjustment model. By using a feedback learning model, key data during task execution is extracted as experience fragments, encoded into genes, and stored in a population gene bank. Then, through a gene evolution model, the experience fragments in the population gene bank are screened, cross-linked, and mutated, and the evolved experience fragments are synchronously updated to all humanoid robots.

2. The method according to claim 1, characterized in that, Based on the task node, the multi-robot collaboration and task allocation model, according to the current state and task priority of each humanoid robot, performs task bidding and dynamic allocation among multiple humanoid robots, including: Publish the multiple task nodes to the task marketplace; Each humanoid robot is controlled to generate candidate execution plans for task nodes in the task market and bid on them based on its current position, load status, tool availability, and task execution capabilities. The multi-robot collaboration and task allocation model comprehensively evaluates the bidding schemes of each legged humanoid robot, task priorities, and the principle of avoiding path conflicts, so as to dynamically allocate the task nodes to the most suitable humanoid robot.

3. The method according to claim 2, characterized in that, The control system, which involves each humanoid robot generating candidate execution plans for task nodes in the task market based on its current position, load status, tool availability, and task execution capabilities, and then bidding on these plans, includes: Control each humanoid robot to scan the surrounding environment in real time to obtain environmental perception data; Based on the environmental perception data and the requirements of the task nodes, one or more candidate execution schemes are evaluated and generated; the candidate execution schemes include at least the estimated path planning, gait strategy and task completion time.

4. The method according to claim 1, characterized in that, Based on the environmental model and the task nodes assigned to the current humanoid robot, the model predictive control algorithm is used for multi-objective optimization to generate an optimal control strategy that includes the walking path and gait, including: Construct a prediction model based on the aforementioned humanoid robot dynamics model; Define a multi-objective cost function that includes path tracking error, gait stability, energy consumption, and penalty terms related to task priority; In each control cycle, with the current humanoid robot state and the environmental model as initial conditions, the optimal control input sequence for a future period is obtained by solving the minimization problem of the multi-objective cost function within the set prediction time domain. The first control input in the optimal control input sequence is transmitted to the current humanoid robot, and the steps are repeated in the next control cycle. In each control cycle, with the current humanoid robot state and the environment model as initial conditions, the optimal control input sequence for a future period is obtained by solving the minimization problem of the multi-objective cost function within the set prediction time domain, so as to calculate the optimal control strategy including walking path and gait.

5. The method according to claim 4, characterized in that, The penalty term related to task priority in the multi-objective cost function is configured such that for high-priority task nodes, the path tracking error term corresponding to the high-priority task node has a weight coefficient higher than a preset weight threshold, so as to drive the humanoid robot to complete high-priority tasks first.

6. The method according to claim 1, characterized in that, The process of controlling the current legged humanoid robot to execute the optimal control strategy, and dynamically adjusting the optimal control strategy according to the environment model through a task execution dynamic adjustment model, includes: If a new obstacle or a change in the position of the task target is detected during the humanoid robot's task execution, path replanning is immediately triggered to update the environment model, resulting in an updated environment model. Based on the updated environment model, the optimal control strategy is recalculated through the task execution dynamic adjustment model. If a task priority change instruction is received from the manufacturing execution system, the order or path planning of the currently executed tasks will be dynamically adjusted according to the changed task priority.

7. The method according to any one of claims 1 to 6, characterized in that, The process involves using a gene evolution model to screen, crossover, and mutate experience fragments in the population's gene pool, and then synchronously updating these evolved experience fragments to all humanoid robots, including: Based on the success rate, execution efficiency, or energy consumption indicators of historical tasks, the fitness of the empirical fragments in the population gene pool is evaluated. Based on the fitness evaluation results, experience fragments with high fitness are selected for retention; The selected experience fragments are subjected to crossover and mutation operations to generate the evolved experience fragments and update them to all humanoid robots.

8. A walking control device for a humanoid robot, characterized in that, The device includes: The task processing module is used to receive production work orders from the manufacturing execution system, parse the production work orders to obtain multiple task nodes, and based on the task nodes, through a multi-robot collaboration and task allocation model, to conduct task bidding and dynamic allocation among multiple humanoid robots according to the current status and task priority of each humanoid robot. The environment construction module is used to acquire real-time environmental data and construct an environment model that includes ground conditions, obstacles, the robot's own state, and the task target area based on the real-time environmental data. The strategy optimization module is used to perform multi-objective optimization based on the environment model and the task nodes assigned to the current humanoid robot, using a model predictive control algorithm to generate an optimal control strategy that includes walking path and gait; wherein, the multi-objective optimization considers at least task priority and environmental uncertainty; The strategy adjustment module is used to control the current legged humanoid robot to execute the optimal control strategy, and to dynamically adjust the optimal control strategy according to the environment model through the task execution dynamic adjustment model; The experience evolution module is used to extract key data from the task execution process into experience fragments through a feedback learning model, encode them into genes and store them in a population gene bank. The gene evolution model is used to screen, crossover and mutate the experience fragments in the population gene bank and synchronously update the evolved experience fragments to all humanoid robots.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images