Control system of cloud collaborative humanoid robot

Abstract

The invention discloses a cloud collaborative robot control system, which comprises a cloud control platform and a robot, and is characterized in that the robot is provided with an edge calculation module and a hardware module; the hardware module provides hardware components, collects environment data and uploads the environment data to the cloud control platform. The edge computing module monitors the state of the robot, collects execution data, uploads the execution data to the cloud control platform, updates local model parameters of the robot, uploads the parameters to the cloud control platform, and controls the robot according to a task strategy and an optimization strategy; the cloud control platform formulates a task strategy according to the received production task and environmental data based on the large model, formulates an optimization strategy according to the received uploaded data, updates the large model according to the uploaded data, local model parameters and a task experience pool, and issues the task strategy, the optimization strategy and large model parameters to the robot; and the task execution efficiency and the system control efficiency are improved. The embodiment of the invention can be widely applied to the technical field of robot control.

Description

Technical Field

[0001] The present invention relates to the field of robot control technology, and in particular to a cloud-based collaborative humanoid robot control system. Background Art

[0002] With the development of intelligent manufacturing, more and more robots are involved in production tasks; the existing multi-robot scheduling system usually adopts preset scheduling algorithms and centralized control modes, which are difficult to cope with changes in the environment and changes in task requirements, and have low task execution efficiency and collaborative control efficiency. Summary of the invention

[0003] The main purpose of the embodiment of the present invention is to provide a control system for a cloud-based collaborative humanoid robot, which can improve task execution efficiency and collaborative control efficiency.

[0004] To achieve the above-mentioned purpose, an embodiment of the present invention provides a control system of a cloud-cooperative humanoid robot, the system comprising a cloud-cooperative control platform and a humanoid robot, the humanoid robot comprising an edge computing module and a hardware module; wherein:

[0005] The hardware module is used to collect environmental data and send the environmental data to the cloud control platform via the edge computing module; provide hardware components so that the robot can perform production tasks through the hardware components;

[0006] The edge computing module is used to monitor the status of the humanoid robot, obtain monitoring data, and send it to the cloud control platform; control the humanoid robot to perform the production task according to the task strategy and optimization strategy, and feed back the execution data to the cloud control platform; update the local model parameters according to the model parameters, obtain local parameters, and send the local parameters to the cloud control platform;

[0007] The cloud control platform is used to receive the production task, input the production task and the environmental data into the big model for analysis, and determine the task strategy; determine the optimization strategy according to the execution data, the monitoring data and the environmental data; update the big model according to the environmental data, the local parameters and the task experience pool to obtain the model parameters; and send the task strategy, the optimization strategy and the model parameters to the humanoid robot.

[0008] In some embodiments, the cloud control platform includes any combination of an intelligent decision-making system, an intelligent distribution system, an intelligent energy management system, an adaptive software system, or an experience sharing system; wherein,

[0009] The intelligent decision-making system is used to perform deep learning based on the production task, the environmental data and the large model to obtain the task strategy; perform processing based on the execution data, the environmental data and the large model to obtain the optimization strategy; analyze the monitoring data, and if there is abnormal state data, analyze the abnormal state data to determine a solution; generate a fault report based on the abnormal data and the solution;

[0010] The intelligent allocation system is used to analyze the production tasks and the monitoring data to determine the task allocation plan; analyze the execution data and the monitoring data to determine the resource sharing plan;

[0011] The intelligent energy management system is used to analyze the monitoring data, determine the energy consumption data, perform energy consumption analysis based on the production tasks, the energy consumption data and the large model, and determine the energy consumption optimization plan; analyze the execution data and the monitoring data to determine the power distribution plan;

[0012] The adaptive software system is used to generate a custom production task in response to external input; and generate a feedback report based on the execution data and the monitoring data;

[0013] The experience sharing system is used to build a task experience pool, update the task experience pool according to the execution data; update the parameters of the large model according to the task experience pool to obtain model parameters; and optimize the parameters of the large model according to the local parameters to obtain the optimized large model.

[0014] In some embodiments, the intelligent decision-making system includes a strategy generation module, a strategy optimization module and a fault handling module; wherein,

[0015] The strategy generation module is used to analyze the production task to obtain task characteristics, analyze the environmental data to obtain environmental characteristics; match the task characteristics and the environmental characteristics with the task experience pool to determine the first solution; input the task characteristics and the environmental characteristics into the large model for analysis to obtain the task type and environmental factors; process the task type, the environmental factors, the first solution and the large model to obtain the task strategy;

[0016] The strategy optimization module is used to analyze the execution data to determine the task progress data, analyze the environment data to determine the environment change data; and analyze according to the task progress data, the environment change data, the large model and the preset adjustment strategy to obtain the optimization strategy;

[0017] The fault processing module is used to extract data from the monitoring data to obtain robot status data; identify abnormal events based on the large model and the robot status data; generate warning information based on the abnormal events, determine the abnormality type based on the abnormal events, and determine the solution based on the abnormality type and a preset processing strategy.

[0018] In some embodiments, the intelligent allocation system includes a task allocation module and a resource sharing module; wherein,

[0019] The task allocation module is used to analyze the production task and determine the task priority; analyze the monitoring data to determine the robot status data, and determine the fitness score according to the relationship between the robot status data and the preset score; match the task priority and the fitness score to determine the task allocation plan;

[0020] The resource sharing module is used to analyze the monitoring data to determine resource usage data; analyze the execution data to determine resource requirements; determine a resource sharing request based on the resource usage data and the resource requirements; and determine the resource sharing solution based on the resource sharing request and the resource usage.

[0021] In some embodiments, the intelligent energy management system includes an energy consumption optimization module and a power management module; wherein,

[0022] The energy consumption optimization module is used to extract the monitoring data to obtain energy consumption data; analyze the production tasks to determine the task priorities; input the energy consumption data and the task priorities into the large model for processing to obtain an energy consumption optimization strategy;

[0023] The power management module is used to analyze the execution data, determine the task load data, and determine the load level according to the task load data and a preset threshold; extract data from the monitoring data to obtain power monitoring data; determine a power distribution plan according to the load level and the power monitoring data; compare the power monitoring data with a preset threshold, and if the power monitoring data is less than or equal to the preset threshold, switch to a backup power supply and charge the power supply before switching.

[0024] In some embodiments, the adaptive software system includes a task customization module and an interactive feedback module; wherein,

[0025] The task customization module is used to analyze the external input request and determine the task mode; generate a customized production task according to the task mode; wherein the task mode includes task type, module configuration, execution parameters, resource allocation strategy and fault response strategy;

[0026] The interactive feedback module is used to generate a feedback report based on the monitoring data and the execution data, and send the feedback report, the monitoring data and the execution data to the human-computer interaction interface so that the human-computer interaction interface is displayed based on the feedback report, the monitoring data and the execution data.

[0027] In some embodiments, the experience sharing system includes a task experience library and a model updating module; wherein,

[0028] The task experience library is used to record the execution data;

[0029] The model updating module is used to classify the task experience library, update the large model according to the classified task experience library, and obtain model parameters; aggregate the acquired local parameters to obtain the first parameters, and optimize the large model according to the first parameters to obtain the optimized large model.

[0030] In some embodiments, the hardware module includes any combination of a fixture module, a mobile module, a sensor module, or an arm and joint module; wherein,

[0031] The fixture module, the mobile module, and the arm and joint module are used to provide replaceable hardware components for the humanoid robot;

[0032] The sensing module is used to collect data on the current environment of the humanoid robot, obtain the environmental data, and send the environmental data to the edge computing module.

[0033] In some embodiments, the edge computing module is also used to receive the environmental data and upload the environmental data to the cloud control platform; analyze the monitoring data, determine the task progress data, and upload the task progress data to the cloud control platform; parse the task strategy or the optimization strategy, extract control instructions, and control the humanoid robot according to the control instructions; monitor the power data of the humanoid robot, compare the power data with a preset value, and if the power data is less than or equal to the preset value, control the humanoid robot to switch to energy-saving mode.

[0034] In some embodiments, the edge computing module is also used to update the local model of the humanoid robot using the model parameters to obtain an updated local model; obtain the current environment data and current execution data of the humanoid robot, and adjust the parameters of the updated local model according to the current environment data and the current execution data to obtain updated local model parameters; encrypt the updated local model parameters to obtain encrypted data, and upload the encrypted data to the cloud control platform.

[0035] The implementation of the embodiments of the present invention includes the following beneficial effects: the present embodiment provides a control system for a cloud-cooperative humanoid robot, the system includes a cloud control platform and a humanoid robot, the humanoid robot includes an edge computing module and a hardware module; the hardware module collects environmental data, and sends the environmental data to the cloud control platform via the edge computing module, and provides hardware components for the humanoid robot to perform production tasks; the edge computing module monitors the status of the humanoid robot, and feeds back the status of the humanoid robot to the cloud control platform, controls the humanoid robot to perform production tasks according to task strategies and optimization strategies, feeds back execution data to the cloud control platform, updates the local model according to the model parameters, and sends the updated model parameters to the cloud control platform; the cloud control platform receives the production task , analyze the production tasks and environmental data according to the big model to obtain the task strategy; determine the optimization strategy according to the execution data, monitoring data and environmental data, update the big model according to the environmental data, local parameters and task experience pool, and send the obtained task strategy, optimization strategy and parameters of the big model to the humanoid robot; formulate task strategies and optimization strategies according to the collected data through the edge computing module and the cloud control platform, control the humanoid robot to perform production tasks, realize adaptive task scheduling and optimization, and improve the task execution efficiency and collaborative control efficiency; set up the task experience pool, edge computing module and big model for experience sharing, optimize the big model parameters, improve the efficiency of formulating task strategies and optimization strategies, and thus improve the task execution efficiency and collaborative control efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic diagram of an implementation environment of a control system of a cloud-based collaborative humanoid robot provided by an embodiment of the present invention;

[0037] Figure 2 It is a structural block diagram of a control system of a cloud-based collaborative humanoid robot provided by an embodiment of the present invention;

[0038] Figure 3 It is a structural block diagram of a cloud control platform in a control system of a cloud collaborative humanoid robot provided by an embodiment of the present invention;

[0039] Figure 4It is a structural block diagram of an intelligent decision-making system of a cloud control platform in a control system of a cloud-based collaborative humanoid robot provided by an embodiment of the present invention;

[0040] Figure 5 It is a structural block diagram of an intelligent distribution system of a cloud control platform in a control system of a cloud collaborative humanoid robot provided by an embodiment of the present invention;

[0041] Figure 6 It is a structural block diagram of an intelligent energy management system of a cloud control platform in a control system of a cloud-based collaborative humanoid robot provided by an embodiment of the present invention;

[0042] Figure 7 It is a structural block diagram of an adaptive software system of a cloud control platform in a control system of a cloud collaborative humanoid robot provided by an embodiment of the present invention;

[0043] Figure 8 It is a structural block diagram of an experience sharing system of a cloud control platform in a control system of a cloud collaborative humanoid robot provided by an embodiment of the present invention;

[0044] Fig. 9 It is a schematic diagram of the step flow of a specific embodiment provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only provided for the convenience of explanation and description, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0046] In the following description, reference is made to “some embodiments”, which describe a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0047] In the following description, the terms "first\second\third" involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that "first\second\third" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present invention described herein can be implemented in an order other than that illustrated or described herein.

[0048] Unless otherwise defined, all technical and scientific terms used in the embodiments of the present invention have the same meanings as those commonly understood by those skilled in the art of the present invention. The terms used in the embodiments of the present invention are only for the purpose of describing the embodiments of the present invention and are not intended to limit the present invention.

[0049] See also Figure 1 , Figure 1 It is an implementation environment of a control system of a cloud-based collaborative humanoid robot provided by an embodiment of the present invention; a cloud-based control platform is set in the cloud, and the cloud-based control platform can be connected to several humanoid robots. The humanoid robots have edge computing functions, so that the humanoid robots have a certain level of intelligence.

[0050] See also Figure 2 , Figure 2 : is a structural block diagram of a control system of a cloud-cooperative humanoid robot provided by an embodiment of the present invention, the system includes a cloud control platform and a humanoid robot, the humanoid robot includes an edge computing module and a hardware module; wherein,

[0051] The hardware module is used to collect environmental data and send the environmental data to the cloud control platform through the edge computing module; provide hardware components so that the humanoid robot can perform production tasks through the hardware components;

[0052] The edge computing module is used to monitor the status of the humanoid robot, obtain monitoring data, and send it to the cloud control platform; control the humanoid robot to perform production tasks according to the task strategy and optimization strategy, and feed back the execution data to the cloud control platform; update the local model parameters according to the model parameters, obtain local parameters, and send the local parameters to the cloud control platform;

[0053] The cloud control platform is used to receive production tasks, input production tasks and environmental data into the big model for analysis, and determine the task strategy; determine the optimization strategy based on the execution data, monitoring data and environmental data; update the big model based on the environmental data, local parameters and task experience pool to obtain model parameters; and send the task strategy, optimization strategy and model parameters to the humanoid robot.

[0054] In some embodiments, the hardware module is a part of the humanoid robot, which provides the humanoid robot with a hardware module for performing production tasks, including a fixture module, a mobile module, a sensor module, and an arm and joint module; each humanoid robot is provided with an edge computing module to process the collected data and the received data for the humanoid robot, and update the local model of the humanoid robot to reduce data delay and improve the response speed and intelligence level of the system. The cloud control platform is set up in the cloud, receives data uploaded by all humanoid robots connected to the cloud control platform, processes and analyzes the uploaded data, formulates relevant strategies, and generates corresponding control instructions to control all humanoid robots to work together to complete production tasks; the cloud control platform is provided with a large model and a task experience pool to record the data of the humanoid robot performing tasks, and continuously optimizes the large model through collective learning and experience sharing, thereby improving the intelligence level of the entire system and the efficiency of task execution; each humanoid robot is connected to the cloud control platform through wireless communication, such as Bluetooth, 5G, etc., but not limited to this.

[0055] See also Figure 3 In some embodiments, the cloud control platform includes an intelligent decision-making system, an intelligent distribution system, an intelligent energy management system, an adaptive software system and an experience sharing system;

[0056] Intelligent decision-making system, which is used to perform deep learning based on production tasks, environmental data and large models to obtain task strategies; process execution data, environmental data and large models to obtain optimization strategies; analyze monitoring data, and if there is abnormal status data, analyze the abnormal status data and determine solutions; generate fault reports based on abnormal data and solutions;

[0057] Intelligent allocation system, which is used to analyze production tasks and monitoring data to determine task allocation plans; analyze execution data and monitoring data to determine resource sharing plans;

[0058] Intelligent energy management system, used to analyze monitoring data, determine energy consumption data, perform energy consumption analysis based on production tasks, energy consumption data and large models, and determine energy consumption optimization plans; analyze execution data and monitoring data to determine power allocation plans;

[0059] Adaptive software system for generating customized production tasks in response to external inputs; generating feedback reports based on execution and monitoring data;

[0060] The experience sharing system is used to build a task experience pool and update the task experience pool according to the execution data; update the parameters of the large model according to the task experience pool to obtain the model parameters; optimize the parameters of the large model according to the local parameters to obtain the optimized large model.

[0061] In some embodiments, the cloud control platform is mainly responsible for advanced management functions such as intelligent decision-making, task scheduling, experience sharing and collective learning. It is also responsible for handling edge computing synchronization and coordination, generating execution strategies and optimization strategies for production tasks, and issuing them to the humanoid robots connected to the cloud control platform; therefore, the cloud control platform is provided with an intelligent decision-making system, an intelligent distribution system, an intelligent energy management system and an experience sharing system according to the functions it is responsible for; in this embodiment, the cloud control platform has user-defined functions, and users can customize the content of the tasks to be performed; therefore, the cloud control platform is also provided with a custom software system; by setting up an intelligent decision-making system, the cloud control platform uses the set large model to process the data uploaded by each humanoid robot, and generates the execution strategies and optimization strategies for the corresponding production tasks; the generated execution strategies and optimization strategies are issued to the appropriate people through the intelligent distribution system. The cloud control platform collects data from the humanoid robots during the production task execution in real time, records and saves them in the experience sharing system, and uses the recorded data to continuously optimize the cloud control platform's big model based on group learning and distributed learning, thereby improving the accuracy and computational efficiency of the big model and thus improving the efficiency of the humanoid robots in executing tasks. The cloud control platform also realizes human-computer interaction through a custom software system, generates user-defined production tasks, and adjusts the execution parameters of the system in real time through a custom software system to improve execution efficiency.

[0062] See also Figure 4 In some embodiments, the intelligent decision-making system includes a strategy generation module, a strategy optimization module and a fault handling module; wherein,

[0063] The strategy generation module is used to analyze the production task to obtain the task characteristics, analyze the environmental data to obtain the environmental characteristics; match the task characteristics and environmental characteristics with the task experience pool to determine the first plan; input the task characteristics and environmental characteristics into the big model for analysis to obtain the task type and environmental factors; process the task type, environmental factors, the first plan and the big model to obtain the task strategy;

[0064] The strategy optimization module is used to analyze the execution data to determine the task progress data, analyze the environmental data to determine the environmental change data; analyze the task progress data, environmental change data, large model and preset adjustment strategy to obtain the optimization strategy;

[0065] The fault handling module is used to extract monitoring data and obtain robot status data; identify abnormal events based on the large model and robot status data; generate warning information based on abnormal events, determine the abnormal type based on the abnormal event, and determine the solution based on the abnormal type and preset processing strategy.

[0066] In some embodiments, the intelligent decision-making system makes decisions through a large model and a task experience pool set up on a cloud control platform. The intelligent decision-making system is provided with a strategy generation module, a strategy optimization module and a fault handling module; the strategy generation module uses the large model to conduct an in-depth analysis of the received production tasks and the environmental data collected by the humanoid robot to determine the task characteristics such as the task type and task requirements of the production task, as well as the current position of the humanoid robot, the status of each device of the humanoid robot, the status of the hardware resources and other environmental characteristics; based on the analyzed task characteristics and environmental characteristics, a preliminary match is performed in the task experience pool to determine whether there is a task strategy corresponding to similar characteristics. If so, it can be based on the existing task strategy If the robot is not sure about the production process, it can make slight adjustments to improve the production efficiency. If not, it can plan through the big model and generate the task strategy of the production task to control the robot to perform the corresponding production task. The strategy optimization module uses the big model to generate the optimization strategy based on the task execution progress, environmental change data and other information fed back by the robot in real time. When the robot encounters obstacles or deviations, it will make adjustments according to the optimization strategy to improve the task execution efficiency. The fault handling module monitors the status of the cloud control platform, robots and equipment in real time, identifies whether a fault occurs through the big model and the task experience pool, and generates corresponding solutions when a fault occurs, and sends out alarm information to improve the overall fault tolerance of the system and the stability of system operation.

[0067] See also Figure 5 In some embodiments, the intelligent allocation system includes a task allocation module and a resource sharing module; wherein,

[0068] The task allocation module is used to analyze production tasks and determine task priorities; analyze monitoring data to determine robot status data, and determine fitness scores based on the relationship between robot status data and preset scores; match task priorities and fitness scores to determine task allocation plans;

[0069] The resource sharing module is used to analyze the monitoring data to determine the resource usage data; analyze the execution data to determine the resource requirements; determine the resource sharing request based on the resource usage data and resource requirements; and determine the resource sharing plan based on the resource sharing request and resource usage.

[0070] In some embodiments, the cloud control platform allocates production tasks and deploys system resources through an intelligent allocation system; the intelligent allocation system sets up a task allocation module and a resource sharing system, and the task allocation module determines the task requirements of the received production tasks by analyzing, including information such as task type, task urgency, deadline, etc., and assigns priorities to the received production tasks according to the determined task requirements; illustratively, urgent tasks or tasks that need to be executed immediately are assigned high priority, tasks that need to be completed as soon as possible but do not affect the production line are assigned medium priority, and non-urgent tasks that can be executed later are assigned low priority; then, the task allocation module extracts the current status information of each robot from the monitoring data, including the robot's current hardware configuration, current power and load, working status, etc.; based on a preset scoring system, calculates the fitness score according to the acquired task requirements and robot status; and determines the matching robot according to the fitness score, and assigns the production task The task strategy is sent to the matching robot; the resource sharing module analyzes the data uploaded by each robot, obtains the task execution progress, resource usage, etc. of each robot in the system in real time, and synchronizes the obtained data to each robot in real time to achieve data synchronization of the entire system, thereby improving the system's collaborative efficiency and real-time performance; at the same time, the resource sharing module analyzes the uploaded data to determine whether the robot has insufficient resources or excessive load, such as hardware module damage or excessive load, and generates a resource sharing request when insufficient resources or load transition occurs. The uploaded data is analyzed to determine idle related resources, and a resource sharing plan is generated based on the resource sharing request and idle resources, which is sent to the robot. The robot automatically allocates resources according to the issued resource sharing plan, such as the recipient and the sharer collaborate through wireless communication to share idle resources; or the idle robot is controlled through wireless communication to perform part of the tasks for the robot with load transition, thereby improving the resource utilization of the system.

[0071] See also Figure 6 In some embodiments, the intelligent energy management system includes an energy consumption optimization module and a power management module; wherein,

[0072] The energy consumption optimization module is used to extract monitoring data to obtain energy consumption data; analyze production tasks and determine task priorities; input energy consumption data and task priorities into the large model for processing to obtain energy consumption optimization strategies;

[0073] The power management module is used to analyze the execution data, determine the task load data, and determine the load level based on the task load data and the preset threshold; extract the monitoring data to obtain the power monitoring data; determine the power distribution plan based on the load level and the power monitoring data; compare the power monitoring data with the preset threshold. If the power monitoring data is less than or equal to the preset threshold, switch to the backup power supply and charge the power supply before switching.

[0074] In some embodiments, the cloud control platform sets up an intelligent energy management system to allocate and manage the energy of the system. The intelligent energy management system sets up an energy consumption optimization module and a power management module to respectively realize energy consumption optimization and energy management; the energy consumption optimization module obtains the monitoring data of each robot in real time, and extracts the energy consumption information of each robot therefrom, including the power of the robot, the current load, the running modules, etc., and analyzes the extracted energy consumption information through the big model, adjusts the energy consumption allocation and the power supply configuration based on the task load and priority performed by the robot, generates an energy consumption optimization strategy, and sends it to the robot. The robot reduces its own power according to the energy consumption optimization strategy, or optimizes energy consumption, and delivers more resources to the module that performs the task; illustratively, the big model The power consumption data is analyzed to reduce the power of robots performing low-priority tasks or with low loads, thereby reducing unnecessary energy consumption. For robots performing high-priority or high-load tasks, modules not related to the tasks are turned off, and the power of related modules is allocated preferentially to optimize the energy consumption of the robots and ensure the execution of tasks. The power management module monitors the power information and load level of each robot in real time based on the analysis results of the large model, and allocates more power to robots with high load levels to ensure the continuity and efficiency of robot task execution. Robots with low load levels are adjusted to energy-saving mode to reduce energy consumption. The power of the power supply is monitored in real time. When the power is lower than a certain threshold, the backup power supply is switched, and the low-power power supply is charged during the intervals between task executions.

[0075] See also Figure 7 In some embodiments, the adaptive software system includes a task customization module and an interactive feedback module; wherein,

[0076] The task customization module is used to analyze external input requests and determine the task mode; generate customized production tasks according to the task mode; the task mode includes task type, module configuration, execution parameters, resource allocation strategy and fault response strategy;

[0077] The interactive feedback module is used to generate a feedback report based on the monitoring data and the execution data, and send the feedback report, the monitoring data and the execution data to the human-computer interaction interface so that the human-computer interaction interface can display the feedback report, the monitoring data and the execution data.

[0078] In some embodiments, in addition to the production tasks being released by the intelligent production system to the cloud control platform, users can customize the production tasks according to their needs and send them to the cloud control platform through the adaptive software system set up by the cloud control platform; the adaptive software system is provided with a task customization module and an interactive feedback module to realize user-defined functions and human-computer interactive feedback; the task customization module is provided with an open interface, and users can customize the task mode and module configuration according to production needs through the open interface, including custom task types, corresponding process steps, hardware modules required to execute production tasks, execution parameters, etc.; the task customization module determines the corresponding task requirements according to the user-defined content, assigns the corresponding task priority, dynamically adjusts the execution parameters of the system, and matches The corresponding robots execute the production tasks customized by the users, so as to improve the task adaptability of the system. The interactive feedback module is provided with an interactive interface, and the cloud control platform displays the task configuration interface through the interactive interface, so that the users can customize the tasks through the task configuration interface. The interactive feedback module performs statistical analysis on the data uploaded by each robot, and displays the status, task progress, energy consumption and other information of each robot on the interactive interface, so that the users can view and adjust the data in real time. At the same time, the interactive feedback module records the relevant data of the tasks completed by the robots, analyzes and statistics the key data of the task execution, such as the success rate, task time, fault information, etc., and forms a feedback report. The users obtain the feedback report through the interactive interface, and improve the task process and strategy based on the feedback report.

[0079] See also Figure 8 In some embodiments, the experience sharing system includes a task experience library and a model updating module; wherein,

[0080] Task experience database, used to record execution data;

[0081] The model updating module is used to classify the task experience library, update the large model according to the classified task experience library, and obtain model parameters; aggregate the acquired local parameters to obtain the first parameters, and optimize the large model according to the first parameters to obtain the optimized large model.

[0082] In some embodiments, the cloud control platform is provided with a task experience pool to record the execution data uploaded by each robot, including task parameters, environmental conditions, success rate, failure reasons, etc. The task experience pool performs label management on the recorded task execution data, classifies the recorded data, and obtains different types of data, such as different task types, different execution strategies, etc.; the system extracts different types of data from the task experience pool as training samples at a certain period, trains and optimizes the large model set on the cloud control platform, and adjusts the model parameters to make the large model adaptable to different tasks, thereby improving the task execution efficiency; the experience sharing system in the cloud control platform is also provided with a model update module, which can extract sample data from the task experience pool to train the large model, fine-tune the model parameters, realize group learning, and the large model learns the execution experience of different tasks; the model update module sends the updated large model to each robot, and the edge computing module in the robot sends the sent large model As a local model, the edge computing module processes the data and status information collected by the robot through the local model to improve the intelligence level of each robot; after the edge computing module uses the issued large model as the local model, it locally updates the local model according to the robot's current task execution data, environmental data, etc., so that the local model is more adapted to the robot's current environment; the edge computing module adjusts the parameters of the local model in real time according to the current environment and task execution data, and uploads the adjusted local model parameters to the cloud control platform in the set upload cycle. The model update module assigns corresponding weights according to the task completion amount or data amount of different robots. The model update module performs weighted combination of the model parameters uploaded by different robots according to the assigned weights to obtain the global model parameters, and integrates the global model parameters into the large model through the federated learning mechanism to improve the real-time and adaptability of the large model; model decentralization can improve the adaptive adjustment of robots in different production environments.

[0083] In some embodiments, the hardware module of the robot includes any combination of a gripper module, a mobile module, a sensor module, or an arm and joint module; wherein,

[0084] Gripper modules, mobile modules, and arm and joint modules to provide replaceable hardware components for the robot;

[0085] The sensor module is used to collect data about the robot's current environment, obtain environmental data, and send the environmental data to the edge computing module.

[0086] In some embodiments, the hardware module of the robot is a part of the robot, which can be divided into any combination of a fixture module, a mobile module, a sensor module or an arm and joint module according to the application approach. The fixture module sets different fixtures according to the task type of the production task to adapt to workpieces of different materials and shapes, such as magnetic clamps, pneumatic clamps or flexible clamps, so as to improve the flexibility and stability of the robot's grasping; the mobile module is provided with different moving parts according to different production environments, such as wheeled, tracked or mechanical leg modules. The system selects the corresponding mobile module according to the real-time collected environmental data to achieve the optimal movement mode and improve the robot's task execution efficiency; the sensor module integrates a variety of sensors, including visual sensors, lidars, depth cameras and a variety of environmental sensors. The sensor module perceives the robot's environment and task-related workpiece information in real time to generate environmental data; the arm and joint module includes a robotic arm and joint structure with multiple degrees of freedom, so that the robot can perform precise operations and multi-angle task execution to meet the diverse needs of performing different types of tasks.

[0087] In some embodiments, the edge computing module is also used to receive environmental data and upload the environmental data to the cloud control platform; analyze the monitoring data, determine the task progress data, and upload the task progress data to the cloud control platform; parse the task strategy or optimization strategy, extract the control instructions, and control the robot according to the control instructions; monitor the robot's power data, compare the power data with the preset value, and if the power data is less than or equal to the preset value, control the robot to switch to energy-saving mode.

[0088] In some embodiments, the edge computing module set in the robot is also responsible for preprocessing the data collected by the robot and the data released by the cloud control platform, extracting key information from the data, and reducing data transmission delays; at the same time, the edge computing module is also responsible for the communication between the cloud control platform and the robot, improving the real-time response capability of the system; in this embodiment, the edge computing module sends the environmental data collected by the hardware module to the cloud control platform, monitors and analyzes the robot's status and task execution status in real time, determines the task progress of the robot's current production task, and uploads the task progress to the cloud control platform so that the remote control platform can grasp the status information of each robot in real time; the edge computing module parses the task strategy or optimization strategy issued by the cloud control platform, determines the control instructions therein, such as hardware module configuration, planning path and execution parameters, etc., and the edge computing module controls the robot according to the control instructions obtained by the analysis; the edge computing module monitors the robot's power information in real time. When the power information is lower than the preset value, the edge computing module controls the robot to switch to energy-saving mode, or adjusts the execution parameters of the task in real time to reduce energy consumption.

[0089] In some embodiments, the edge computing module is also used to update the model parameters of the robot's local model to obtain an updated local model; obtain the robot's current environment data and current execution data, and adjust the parameters of the updated local model according to the current environment data and current execution data to obtain updated local model parameters; encrypt the updated local model parameters to obtain encrypted data, and upload the encrypted data to the cloud control platform.

[0090] In some embodiments, the edge computing module receives the large model decentralized by the cloud control platform, and improves the intelligence level of the robot based on the decentralized large model, and regularly uploads the local model parameters of the robot to the cloud control platform so that the cloud control platform can update the large model, learn the experience of the large model in performing tasks on different robots, and improve the overall intelligence level of the system; in this embodiment, the edge computing module updates the large model parameters sent by the cloud control platform to the local model of the robot, or uses the large model below as the new local model of the robot. The edge computing module adjusts the parameters of the new local model in real time according to the current environmental data and task execution data of the robot, so as to improve the adaptability of the new local model to the tasks performed by the robot; the edge computing module also encrypts the local model parameters after parameter adjustment according to the preset upload cycle to improve data security, and then uploads the encrypted local model parameters to the cloud control platform to be integrated and updated into the large model, thereby improving the overall intelligence level of the system.

[0091] The following describes the solution of the embodiment of the present invention in detail with reference to specific application examples:

[0092] See also Fig. 9The cloud control platform receives the tasks issued by the intelligent production system. The robot collects environmental data in real time through the vision, lidar and other sensors in the sensor module, including the workpiece position, terrain characteristics and obstacle information, and uploads the collected data to the cloud control platform through wireless communication. The cloud control platform analyzes the task type, target position and working conditions of the production task through the big model. If it is determined that the task is a handling type task, the corresponding fixture module is used to transport the workpiece; the uploaded environmental data is analyzed to determine the terrain, obstacle location and characteristics of the production site. If it is determined that the terrain is flat, a wheeled mobile module can be used to improve the efficiency of task execution; the cloud control platform uses the big model to analyze the results of the analysis. The cloud control platform uses the big model to analyze the task requirements of the production task, including the urgency of the task, the task type and the deadline, and determines that the production task has a high priority. At the same time, the big model is used to analyze the data and production tasks uploaded by the robot edge computing module, including the robot's location, resource status and task requirements, to determine the fitness scores of different robots for the current production task, and to assign the task strategy of the production task to the robot with the highest score according to the task priority and fitness score. The edge computing module controls the robot to execute the production task according to the task strategy, and controls the robot to execute the production task through the sensor module. Upload robot task execution data and environmental change data in real time; the cloud control platform uses the big model to identify the data uploaded by the edge computing module, and detects an obstacle in front of the robot. The system determines that the situation is a medium event through the big model. The big model analyzes the uploaded data and plans the optimization strategy based on the current environmental data and task execution progress, and then sends it to the robot; the robot's edge computing module controls the robot to bypass according to the optimization strategy; the big model updates the task experience library based on the data uploaded by the edge computing module, and the system updates the parameters of the big model based on the task experience library. The system sends the updated big model to all robots, and the robot's edge computing module sends the big model below As a local model; the system uses the large model to analyze the energy consumption of the uploaded data and determines that the current robot's energy consumption is high. The system generates an energy consumption optimization strategy through the large model and sends it to the robot. The robot's edge computing module adjusts the power of each module of the robot according to the energy consumption optimization strategy to optimize the overall energy consumption; the system uses the large model to analyze the uploaded data in real time to determine if the robot has a fault or the robot's visual sensor has a fault. The system uses the large model to analyze the data uploaded by all robots and regulates the robots near the current robot to share idle visual sensors to complete the production task; the system records all task execution data uploaded by the robot, generates a task report, and displays it through an interactive interface.

[0093] The implementation of the embodiments of the present invention includes the following beneficial effects: the present embodiment provides a control system for a cloud-based collaborative robot, the system includes a cloud-based control platform and a robot, the robot includes an edge computing module and a hardware module; the hardware module collects environmental data, and sends the environmental data to the cloud-based control platform via the edge computing module, and provides hardware components for the robot to perform production tasks; the edge computing module monitors the robot status, and feeds back the robot status to the cloud-based control platform, controls the robot to perform production tasks according to task strategies and optimization strategies, feeds back execution data to the cloud-based control platform, updates the local model according to the model parameters, and sends the updated model parameters to the cloud-based control platform; the cloud-based control platform receives the production tasks, and sends the updated model parameters to the cloud-based control platform according to the large model. The model analyzes production tasks and environmental data to obtain task strategies; determines the optimization strategy based on execution data, monitoring data and environmental data, updates the large model based on environmental data, local parameters and task experience pool, and sends the obtained task strategy, optimization strategy and parameters of the large model to the robot; formulates task strategies and optimization strategies based on the collected data through the edge computing module and the cloud control platform, controls the robot to perform production tasks, realizes adaptive task scheduling and optimization, and improves task execution efficiency and collaborative control efficiency; sets up the task experience pool, edge computing module and large model for experience sharing, optimizes the large model parameters, and improves the efficiency of formulating task strategies and optimization strategies, thereby improving task execution efficiency and collaborative control efficiency.

[0094] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A control system for a cloud-based collaborative humanoid robot, characterized in that: The system includes a cloud control platform and a humanoid robot, and the humanoid robot includes an edge computing module and a hardware module; wherein, The hardware module is used to collect environmental data and send the environmental data to the cloud control platform via the edge computing module; provide hardware components so that the humanoid robot can perform production tasks through the hardware components; The edge computing module is used to monitor the status of the humanoid robot, obtain monitoring data, and send it to the cloud control platform; control the humanoid robot to perform the production task according to the task strategy and optimization strategy, and feed back the execution data to the cloud control platform; update the local model parameters according to the model parameters, obtain local parameters, and send the local parameters to the cloud control platform; The cloud control platform is used to receive the production task, input the production task and the environmental data into the big model for analysis, and determine the task strategy; determine the optimization strategy according to the execution data, the monitoring data and the environmental data; update the big model according to the environmental data, the local parameters and the task experience pool to obtain the model parameters; and send the task strategy, the optimization strategy and the model parameters to the humanoid robot.

2. The system according to claim 1, characterized in that The cloud control platform includes any combination of intelligent decision-making system, intelligent distribution system, intelligent energy management system, adaptive software system or experience sharing system; wherein, The intelligent decision-making system is used to perform deep learning based on the production task, the environmental data and the large model to obtain the task strategy; perform processing based on the execution data, the environmental data and the large model to obtain the optimization strategy; analyze the monitoring data, and if there is abnormal state data, analyze the abnormal state data to determine a solution; generate a fault report based on the abnormal data and the solution; The intelligent allocation system is used to analyze the production tasks and the monitoring data to determine the task allocation plan; analyze the execution data and the monitoring data to determine the resource sharing plan; The intelligent energy management system is used to analyze the monitoring data, determine the energy consumption data, perform energy consumption analysis based on the production tasks, the energy consumption data and the large model, and determine the energy consumption optimization plan; analyze the execution data and the monitoring data to determine the power distribution plan; The adaptive software system is used to generate a custom production task in response to external input; and generate a feedback report based on the execution data and the monitoring data; The experience sharing system is used to build a task experience pool, update the task experience pool according to the execution data; update the parameters of the large model according to the task experience pool to obtain model parameters; and optimize the parameters of the large model according to the local parameters to obtain the optimized large model.

3. The system according to claim 2, characterized in that The intelligent decision-making system includes a strategy generation module, a strategy optimization module and a fault handling module; wherein, The strategy generation module is used to analyze the production task to obtain task characteristics, analyze the environmental data to obtain environmental characteristics; match the task characteristics and the environmental characteristics with the task experience pool to determine the first solution; input the task characteristics and the environmental characteristics into the large model for analysis to obtain the task type and environmental factors; process the task type, the environmental factors, the first solution and the large model to obtain the task strategy; The strategy optimization module is used to analyze the execution data to determine the task progress data, analyze the environment data to determine the environment change data; and analyze according to the task progress data, the environment change data, the large model and the preset adjustment strategy to obtain the optimization strategy; The fault processing module is used to extract data from the monitoring data to obtain robot status data; identify abnormal events based on the large model and the robot status data; generate warning information based on the abnormal events, determine the abnormality type based on the abnormal events, and determine the solution based on the abnormality type and a preset processing strategy.

4. The system according to claim 2, characterized in that The intelligent allocation system includes a task allocation module and a resource sharing module; wherein, The task allocation module is used to analyze the production task and determine the task priority; analyze the monitoring data to determine the robot status data, and determine the fitness score according to the relationship between the robot status data and the preset score; match the task priority and the fitness score to determine the task allocation plan; The resource sharing module is used to analyze the monitoring data to determine resource usage data; analyze the execution data to determine resource requirements; determine a resource sharing request based on the resource usage data and the resource requirements; and determine the resource sharing solution based on the resource sharing request and the resource usage.

5. The system according to claim 2, characterized in that The intelligent energy management system includes an energy consumption optimization module and a power management module; wherein, The energy consumption optimization module is used to extract the monitoring data to obtain energy consumption data; analyze the production tasks to determine the task priorities; input the energy consumption data and the task priorities into the large model for processing to obtain an energy consumption optimization strategy; The power management module is used to analyze the execution data, determine the task load data, and determine the load level according to the task load data and a preset threshold; extract data from the monitoring data to obtain power monitoring data; determine a power distribution plan according to the load level and the power monitoring data; compare the power monitoring data with a preset threshold, and if the power monitoring data is less than or equal to the preset threshold, switch to a backup power supply and charge the power supply before switching.

6. The system according to claim 2, characterized in that The adaptive software system includes a task customization module and an interactive feedback module; wherein, The task customization module is used to analyze the external input request and determine the task mode; generate a customized production task according to the task mode; wherein the task mode includes task type, module configuration, execution parameters, resource allocation strategy and fault response strategy; The interactive feedback module is used to generate a feedback report based on the monitoring data and the execution data, and send the feedback report, the monitoring data and the execution data to the human-computer interaction interface so that the human-computer interaction interface is displayed based on the feedback report, the monitoring data and the execution data.

7. The system according to claim 2, characterized in that The experience sharing system includes a task experience library and a model updating module; wherein, The task experience library is used to record the execution data; The model updating module is used to classify the task experience library, update the large model according to the classified task experience library, and obtain model parameters; aggregate the acquired local parameters to obtain the first parameters, and optimize the large model according to the first parameters to obtain the optimized large model.

8. The system according to claim 1, characterized in that The hardware module includes any combination of a fixture module, a mobile module, a sensor module, or an arm and joint module; wherein, The fixture module, the mobile module, and the arm and joint module are used to provide replaceable hardware components for the humanoid robot; The sensing module is used to collect data on the current environment of the humanoid robot, obtain the environmental data, and send the environmental data to the edge computing module.

9. The system according to claim 1, characterized in that The edge computing module is also used to receive the environmental data and upload the environmental data to the cloud control platform; analyze the monitoring data, determine the task progress data, and upload the task progress data to the cloud control platform; parse the task strategy or the optimization strategy, extract the control instructions, and control the humanoid robot according to the control instructions; monitor the power data of the humanoid robot, compare the power data with a preset value, and if the power data is less than or equal to the preset value, control the humanoid robot to switch to energy-saving mode.

10. The system according to claim 1, characterized in that The edge computing module is also used to update the local model of the humanoid robot using the model parameters to obtain an updated local model; obtain the current environment data and current execution data of the humanoid robot, and adjust the parameters of the updated local model according to the current environment data and the current execution data to obtain updated local model parameters; encrypt the updated local model parameters to obtain encrypted data, and upload the encrypted data to the cloud control platform.

Images