Multi-robot intelligent scheduling system based on deep learning

Abstract

The invention provides a multi-robot intelligent scheduling system based on deep learning, and aims to improve the operation efficiency and accuracy of an automatic production line through an advanced algorithm and an Internet of Things technology. The system analyzes historical data by using deep learning, predicts the operation efficiency and accuracy of the robot, and provides a scientific basis for intelligent scheduling decision making. The state information of the robot is collected in real time through the Internet of Things technology, and the real-time performance and accuracy of scheduling are ensured. And the intelligent scheduling module dynamically adjusts task allocation and optimizes the production process in combination with the prediction result and the real-time data. The real-time monitoring interface enables an operator to monitor the state of the robot and adjust the strategy according to feedback. And the analogue simulation module verifies the scheduling scheme before deployment, so that the implementation effect is ensured. According to the system, through an intelligent scheduling strategy, the efficiency of multi-robot collaborative operation is remarkably improved, and innovation is brought to automatic production.

Description

Technical Field

[0001] The present invention relates to the field of automated production technology, and in particular to a multi-robot intelligent scheduling system integrating deep learning, the Internet of Things, cloud computing and artificial intelligence algorithms. Background Art

[0002] In the current field of automated production, with the rapid development of Industry 4.0 and intelligent manufacturing, the application of robotics technology is becoming more and more widespread. The number of robots deployed on the production line is increasing, and they need to work together efficiently when performing complex tasks. However, the existing robot scheduling system often relies on fixed programs and simple logical control, lacking sufficient intelligence and flexibility to cope with changing production needs and environmental changes.

[0003] Traditional scheduling systems often face problems such as unreasonable task allocation, low resource utilization and slow response when dealing with multi-robot collaborative operations. These problems limit the overall efficiency and capacity of the production line, especially when faced with multi-variety, small batch and customized production needs, traditional systems are even more difficult to adapt. In addition, with the expansion of production scale and the diversification of production tasks, manual scheduling and monitoring methods can no longer meet the needs of efficient production, and an intelligent scheduling system is urgently needed to improve the level of automation of operations.

[0004] Therefore, it is particularly important to develop an intelligent scheduling system that can respond to production changes in real time, intelligently optimize task allocation, and coordinate multiple robots to work efficiently. Such a system needs to be able to process a large amount of real-time data, provide decision support through advanced algorithms such as deep learning, and use the Internet of Things technology to achieve interconnection between devices to ensure real-time information transmission and accurate execution of tasks. Through this intelligent scheduling, the operating efficiency and accuracy of the production line can be significantly improved, production costs can be reduced, and the competitiveness of the enterprise can be enhanced. Summary of the invention

[0005] The purpose of the present invention is to provide a multi-robot intelligent scheduling system based on deep learning, which can optimize the collaborative operation of multiple robots in an automated production line through real-time data analysis and intelligent decision-making, and improve production efficiency and operation accuracy. The system realizes the automation and intelligence of the whole process from data collection, processing, analysis to decision-making through a series of highly integrated modules such as deep learning module, Internet of Things module, intelligent scheduling module, real-time monitoring module, simulation module, human-computer interaction module, modular design module, autonomous mobile robot (AMRs) module, cloud computing platform module, etc. The Internet of Things module collects the status information of the robot in real time at a frequency of at least 10Hz, and transmits these data to the data preprocessing unit for cleaning and formatting to ensure that the deep learning model receives high-quality input data. Subsequently, the feature extraction unit extracts key features from the processed data, and the model training unit uses these features to train the deep learning model to predict the operation efficiency and accuracy of the robot. The intelligent scheduling module dynamically adjusts the robot task allocation using genetic algorithm or particle swarm optimization algorithm according to the prediction results of the deep learning model and real-time data. At the same time, the real-time monitoring module updates the monitoring interface of the robot status at a frequency of not less than 5Hz, so that the operator can adjust the scheduling strategy through the human-computer interaction module according to the monitoring information. The simulation module verifies the feasibility of the scheduling strategy in this process and displays the simulation effect through 3D visualization. The modular design module allows the rapid replacement of robot components or modules to adapt to different production tasks. The autonomous mobile robot (AMRs) module performs material handling and transportation tasks, while the cloud computing platform module is responsible for data storage, processing and analysis, and provides remote monitoring and control functions. The communication management unit ensures that the communication between robots is accurate and without delay, the safety monitoring unit is responsible for monitoring the operating status of the robot, and the maintenance management unit tracks the maintenance cycle and needs of the robot to ensure the stable operation of the system. The whole process forms a highly automated and intelligent closed-loop system, realizing the optimization of the entire process from data collection, processing, analysis to decision-making, thereby significantly improving the efficiency and accuracy of multi-robot collaborative operations.

[0006] The main function of the deep learning module is to use historical operation data to train a model that can predict the efficiency and accuracy of robot operations. This module consists of three key units that work together to ensure the efficiency and accuracy of model training.

[0007] First, the data preprocessing unit is responsible for receiving the original historical operation data, which may contain noise and inconsistency. The task of this unit is to clean this data, including removing outliers, filling missing values, and unifying the format, so as to ensure that the data provided to the subsequent processing steps is accurate and usable. Data preprocessing is a very critical step in the machine learning workflow because it directly affects the effect of model training. Next, the feature extraction unit analyzes the preprocessed data to identify and extract key features that can significantly affect the robot's operating efficiency and accuracy. These features may include the robot's operating speed, working time, task type, environmental conditions, etc. The goal of the feature extraction unit is to convert the raw data into a feature set that can provide useful information for the deep learning model, thereby improving the model's prediction ability. Finally, the model training unit uses these extracted features and corresponding labels (i.e., actual operating efficiency and accuracy results) to train the deep learning model. This process involves selecting a suitable deep learning architecture, such as a convolutional neural network or a recurrent neural network, and adjusting model parameters. The goal of the model training unit is to find a model that can accurately predict the robot's operating efficiency and accuracy, with a training accuracy of no less than 90%, which reflects the performance of the model on the training data. Through the work of this module, we can obtain a deep learning model that can accurately predict the robot's operating performance, providing a scientific basis for subsequent intelligent scheduling.

[0008] The IoT module is a key part responsible for data collection in the multi-robot intelligent scheduling system. Its core task is to monitor and collect the operating status of the robot in real time to ensure that the system can make accurate scheduling decisions based on the latest information. The design and technical parameters of the module are as follows: The data collected by the IoT module include but are not limited to the current position, operation speed and task completion of the robot. These parameters are crucial for evaluating the robot's performance and scheduling new tasks. In order to ensure the real-time and accuracy of the data, the collection frequency of the IoT module is set to no less than 10Hz. This means that at least 10 data are collected per second to provide the system with continuous and high-frequency status updates. The IoT module realizes the interconnection and interoperability between production equipment such as robots, sensors, and controllers through standardized communication protocols and interfaces. This connection ensures that data can flow seamlessly between devices and provides a solid data foundation for the intelligent scheduling of the system. In order to further reduce the delay in data transmission, the edge computing function is integrated into the IoT module. The edge computing unit processes data near the source of data generation, which can reduce the distance that data is transmitted in the network, thereby reducing latency. The response time of the edge computing function in the IoT module does not exceed 10ms. This means that the time from data generation to processing completion and availability to the system is extremely short, sufficient to meet the needs of real-time scheduling. In the edge computing unit, the data will be processed as necessary, such as data cleaning, format conversion, feature extraction, etc., to ensure that only useful and well-formatted data is sent to the central processing system.

[0009] The intelligent scheduling module is the core component of the multi-robot intelligent scheduling system. The intelligent scheduling module first integrates the prediction results provided by the deep learning module and the real-time data collected by the Internet of Things module. The deep learning module uses at least 5,000 historical operation samples for training, each sample contains at least 10 feature parameters, and the prediction accuracy is not less than 90%. The data collected by the Internet of Things module includes the current position, operation speed and task completion of the robot, and the collection frequency is not less than 10Hz to ensure the real-time and accuracy of the data. The module uses genetic algorithm or particle swarm optimization algorithm for task allocation. Genetic algorithm simulates natural selection and genetic process, generates new individuals through crossover and mutation operations, and gradually optimizes the solution. The particle swarm optimization algorithm realizes the optimization of the whole group intelligence through communication and experience sharing between particles. The algorithm parameters include population size not less than 50 and number of iterations not less than 100. These parameters ensure the search ability and convergence performance of the algorithm. The intelligent scheduling module dynamically adjusts the task allocation of the robot according to the optimization results of the algorithm. This includes task reallocation, priority adjustment and optimization of job sequence to adapt to changes in production needs. The module can handle multi-task and multi-priority jobs, and achieve efficient scheduling of complex production tasks. The module can respond to changes in the production process in real time, such as changes in robot status, updates on task completion, etc., and quickly re-optimize task allocation. This ensures continuous optimization of production efficiency and job accuracy. The intelligent scheduling module also needs to evaluate the scheduling effect, monitor changes in production efficiency and job accuracy, and feedback and adjust the scheduling strategy to achieve continuous optimization and improvement.

[0010] The real-time monitoring module uses the Internet of Things module to collect the robot's key status information in real time at a frequency of at least 10Hz, including position, speed, task completion, etc. Data is collected through sensors and controllers to ensure the real-time and accuracy of the data. The collected data is transmitted to the central processing system through a wired or wireless network. The security and integrity of data transmission are guaranteed to prevent data loss or damage during transmission. After the data arrives at the central processing system, the real-time monitoring module processes and analyzes the received data. The data is formatted and parsed to extract useful information for subsequent monitoring and analysis. The processed data is updated to the monitoring interface in real time, with an update frequency of not less than 5Hz, to ensure that the operator can see the latest robot status in real time. The interface displays real-time changes in key parameters, such as position tracking, speed changes, etc., as well as immediate alarms for any abnormal conditions. The system automatically records all monitoring operations and event logs of robot operation, and the log data is retained for not less than 30 days. The log can be used to track the historical status of the robot, analyze its performance changes, and troubleshoot when problems occur. The monitoring module has a built-in anomaly detection algorithm that can identify abnormal conditions in the robot's operation in real time, such as speed abnormalities, position deviations, etc. The accuracy of anomaly detection is not less than 95%, ensuring that potential production problems can be discovered and responded to in a timely manner. When an abnormal situation is detected, the system automatically triggers the alarm mechanism and notifies relevant personnel through various means such as interface prompts, sounds, emails, and text messages. Ensure that relevant personnel can take prompt action to reduce the risk of production interruptions. Operators can perform some basic user interaction operations through the monitoring interface, such as viewing detailed logs, configuring alarm thresholds, etc. Provide user-friendly interface design to ensure that operators can easily understand and use the monitoring system. Support remote access function, allowing operators to access the monitoring system from any location via the Internet. With authorization, operators can also remotely execute some control commands, such as adjusting the robot's working parameters or restarting the device.

[0011] The simulation module is a key component of the multi-robot intelligent scheduling system. It uses 3D visualization technology to simulate at least 100 different production scenarios and robot behaviors with millimeter-level accuracy to verify the feasibility of the scheduling strategy proposed by the intelligent scheduling module. The module uses a high-precision physical engine and mathematical model to accurately simulate the robot's movement path, operation accuracy and task execution time, as well as the collaboration and communication between robots, to ensure that the simulation results are highly consistent with the actual production environment. In addition, the simulation module provides a real-time feedback mechanism, allowing operators to easily set simulation parameters, start and stop simulations, and analyze simulation results, so as to test and optimize scheduling strategies in a safe and cost-effective environment, ensuring the effectiveness and reliability of these strategies in actual applications. Through the data interface, the module can also directly import the strategies of the intelligent scheduling module and the real-time data of the Internet of Things module, achieving seamless integration with the production management system and enterprise resource planning system, providing operators with a powerful testing and optimization platform.

[0012] The human-machine interaction module is carefully designed to provide the ultimate user operation experience. It supports interface switching in at least 3 mainstream languages, including but not limited to Chinese, English and Spanish, to accommodate operators with different language backgrounds. The module uses the latest voice recognition technology, which can achieve a recognition accuracy of no less than 98% in noisy industrial environments, allowing operators to quickly and accurately control the robot through voice commands. At the same time, it also integrates an advanced gesture recognition system that can recognize and respond to the operator's gesture commands, as well as a highly sensitive touch screen interface to ensure intuitive operation and fast response. The combination of these functions greatly improves the convenience and intuitiveness of operation, reduces the difficulty of operation, and improves production efficiency.

[0013] The modular design module is a key technology to improve the adaptability and flexibility of the production line. It allows operators to quickly replace or upgrade key components or modules of the robot, such as the end effector, sensor or control unit, within 2 hours to adapt to different production tasks or technology updates. This modular design reduces the time required for equipment conversion and maintenance, improves production efficiency, and reduces downtime caused by equipment adjustments. In addition, the modular design also supports rapid fault diagnosis and repair, because the module can be tested and repaired independently of the main system, thereby reducing the downtime of the production line.

[0014] The autonomous mobile robot (AMRs) module is designed to improve logistics efficiency. It is capable of handling and transporting materials at a speed of no less than 2m / s in a factory environment, ensuring rapid material flow and reducing production delays. The module integrates advanced path planning algorithms that can optimize transportation routes in real time to avoid congestion and conflicts. At the same time, the obstacle avoidance algorithm can respond within 0.5 seconds, ensuring that the path is quickly adjusted when an obstacle is detected to avoid collisions and ensure the safety of personnel and equipment. The combination of these functions enables the AMRs module to improve logistics efficiency while ensuring the safety and reliability of the material handling process.

[0015] The cloud computing platform module provides the system with powerful data processing and storage capabilities. It supports data processing capabilities of no less than 1TB / s, and can process a large amount of real-time data from the IoT module to ensure real-time data analysis and decision support. At the same time, the storage capacity provided by the module is no less than 100TB, which provides the system with sufficient data storage space, supports the storage and analysis of historical data, and the training and optimization of machine learning models. In addition, the cloud computing platform also supports remote monitoring and control functions, allowing operators to access the system from any location through a secure network connection, monitor production status in real time, and remotely adjust production parameters to ensure efficient operation and scalability of the system.

[0016] The system operation steps are as follows:

[0017] S1: The system starts, initializes, and completes self-check to ensure that all modules are operating normally.

[0018] S2: Data collection, the IoT module starts to collect the robot's status information in real time at a frequency of at least 10 Hz.

[0019] S3: Data processing, the data preprocessing unit cleans and formats the collected data, and the feature extraction unit extracts key features from it.

[0020] S4: Deep learning prediction, the model training unit uses the extracted features to train the deep learning model to predict the robot's operating efficiency and accuracy.

[0021] S5: Intelligent Scheduling,The intelligent scheduling module dynamically adjusts the robot task allocation using genetic algorithm or particle swarm optimization algorithm based on the prediction results of the deep learning model and real-time data.

[0022] S6: Simulation verification, the simulation module verifies the feasibility of the scheduling strategy and simulates different production scenarios and robot behaviors.

[0023] S7: Task execution. According to the verified scheduling strategy, the robot starts to execute the assigned task.

[0024] S8: Real-time monitoring and exception handling. The real-time monitoring module updates the monitoring interface of the robot status at a frequency of not less than 5Hz. The exception detection function identifies and responds to abnormal situations in the production process in real time.

[0025] S9: Human-machine interaction, the operator interacts with the robot through the human-machine interaction module to adjust the scheduling strategy or perform remote control.

[0026] S10: System optimization: The system optimizes itself based on real-time feedback and monitoring data, updates scheduling strategies, and ensures continuous efficiency and accuracy improvements.

[0027] Through the collaborative work of the above modules, the multi-robot intelligent scheduling system of the present invention can realize the efficient collaborative operation of robots, improve the operating efficiency and accuracy of the production line, and meet the needs of modern manufacturing industry for flexible and efficient production processes.

[0028] By adopting the technical solution of the present invention, the multi-robot intelligent scheduling system can bring the following clear technical effects:

[0029] 1. Improve production efficiency: The system analyzes historical operation data through the deep learning module and predicts the operating efficiency and accuracy of each robot, with an accuracy rate of no less than 90%. The intelligent scheduling module uses these prediction results and the robot status information collected in real time by the IoT module to dynamically optimize task allocation to ensure that each robot can work on the most suitable task. This real-time optimization can reduce machine idle and waiting time by at least 20%, significantly improving the overall operating efficiency of the production line.

[0030] 2. Enhanced operation accuracy and reliability: The real-time monitoring module updates the robot status information at a frequency of no less than 5Hz, and combined with the anomaly detection function, it can identify abnormal situations in the production process in real time with an accuracy of no less than 95%. Once an abnormality is detected, the system will immediately notify the operator so that he can take prompt measures to avoid production accidents. In addition, the simulation module can simulate at least 100 different production scenarios, verify the effectiveness of the scheduling strategy with millimeter-level accuracy, and correct possible problems in advance, thereby improving the accuracy and reliability of operations.

[0031] 3. Improve the flexibility and adaptability of the system: The modular design module allows for rapid replacement of robot parts or modules within no more than 2 hours, allowing the production line to quickly adapt to different production tasks. The transportation speed of the autonomous mobile robot (AMRs) module is not less than 2m / s, and the obstacle avoidance response time is no more than 0.5 seconds, ensuring high efficiency and safety of material handling. The human-computer interaction module supports at least 3 language interface switching, and the voice recognition accuracy is not less than 98%, allowing operators to operate and manage more intuitively and conveniently, improving the system's adaptability to different production environments and operating habits. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a block diagram of the multi-robot intelligent scheduling system of the present invention.

[0033] Figure 2 This is a diagram of the operation steps of the multi-robot intelligent scheduling system of the present invention. DETAILED DESCRIPTION

[0034] The following will be combined with the attached embodiment of the present invention Figure 1 and Figure 2 , the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] The purpose of the present invention is to provide a multi-robot intelligent scheduling system based on deep learning, which can optimize the collaborative operation of multiple robots in an automated production line through real-time data analysis and intelligent decision-making, and improve production efficiency and operation accuracy. The system realizes the automation and intelligence of the whole process from data collection, processing, analysis to decision-making through a series of highly integrated modules such as deep learning module, Internet of Things module, intelligent scheduling module, real-time monitoring module, simulation module, human-computer interaction module, modular design module, autonomous mobile robot (AMRs) module, and cloud computing platform module.

[0036] The main function of the deep learning module is to use at least 5,000 historical operation sample data to train a model that can predict the efficiency and accuracy of the robot's operation. Each of these samples contains at least 10 feature parameters, such as the robot's movement speed, operation accuracy, task complexity, environmental variables, etc. The module consists of three key units that work together to ensure the efficiency and accuracy of model training: Data preprocessing unit, which is responsible for receiving the original historical operation data, which may contain noise and inconsistency. The task of this unit is to clean this data, including removing outliers, filling missing values, and unifying the format, so as to ensure that the data provided to the subsequent processing steps is accurate and usable. The data preprocessing unit ensures that the data cleaning accuracy is not less than 99%. The feature extraction unit analyzes the preprocessed data to identify and extract key features that can significantly affect the robot's operation efficiency and accuracy. These features may include the robot's running speed, working time, task type, environmental conditions, etc. The goal of the feature extraction unit is to convert the raw data into a feature set that can provide useful information for the deep learning model, thereby improving the model's predictive ability. The feature extraction accuracy is not less than 95%. The model training unit uses these extracted features and corresponding labels (i.e., actual operating efficiency and accuracy results) to train a deep learning model. This process involves selecting an appropriate deep learning architecture, such as a convolutional neural network or a recurrent neural network, and adjusting model parameters. The goal of the model training unit is to find a model that can accurately predict the operating efficiency and accuracy of the robot, with a training accuracy of no less than 90%, which reflects the performance of the model on the training data.

[0037] The IoT module monitors and collects the operating status of the robot in real time to ensure that the system can make accurate scheduling decisions based on the latest information. The design and technical parameters of the module are as follows: The data collected by the IoT module include but are not limited to the current position, operation speed and task completion of the robot. These parameters are essential for evaluating the robot's performance and scheduling new tasks. In order to ensure the real-time and accuracy of the data, the collection frequency of the IoT module is set to no less than 10Hz. This means that data is collected at least 10 times per second to provide the system with continuous and high-frequency status updates. The IoT module realizes the interconnection and interoperability between production equipment such as robots, sensors, and controllers through standardized communication protocols and interfaces. This connection ensures that data can flow seamlessly between devices and provides a solid data foundation for the intelligent scheduling of the system. In order to further reduce the delay of data transmission, edge computing functions are integrated into the IoT module. The edge computing unit processes data near the source of data generation, which can reduce the distance that data is transmitted in the network, thereby reducing latency. The response time of the edge computing function in the IoT module does not exceed 10ms. This means that the time from data generation to processing completion and availability to the system is extremely short, sufficient to meet the needs of real-time scheduling. In the edge computing unit, the data will be processed as necessary, such as data cleaning, format conversion, feature extraction, etc., to ensure that only useful and well-formatted data is sent to the central processing system.

[0038] The intelligent scheduling module integrates the prediction results of the deep learning module and the real-time data collected by the Internet of Things module. The deep learning module uses at least 5,000 historical job samples for training, each sample contains at least 10 feature parameters, and the prediction accuracy is not less than 90%. The data collected by the Internet of Things module includes the current position, operation speed and task completion status of the robot, and the collection frequency is not less than 10Hz to ensure the real-time and accuracy of the data. This module uses genetic algorithm or particle swarm optimization algorithm for task allocation. The parameters of the genetic algorithm include population size not less than 50, termination evolution generation not less than 100, crossover probability between 0.4 and 0.99, and mutation probability between 0.0001 and 0.1. The parameters of the particle swarm optimization algorithm mainly include the number of particles not less than 50, the maximum speed determines the search step between the current position of the particle and the optimal position, individual learning factor, social learning factor and inertia factor. The reasonable setting of these parameters plays a vital role in the performance of the algorithm. The intelligent scheduling module dynamically adjusts the task allocation of the robot according to the optimization results of the algorithm. This includes task reallocation, priority adjustment and optimization of job sequence to adapt to changes in production needs. The module can handle multi-task and multi-priority jobs, and achieve efficient scheduling of complex production tasks. The module can respond to changes in the production process in real time, such as changes in robot status, updates on task completion, etc., and quickly re-optimize task allocation. This ensures continuous optimization of production efficiency and job accuracy. The intelligent scheduling module also needs to evaluate the scheduling effect, monitor changes in production efficiency and job accuracy, and feedback and adjust the scheduling strategy to achieve continuous optimization and improvement.

[0039] The real-time monitoring module collects the robot's status information, including position, speed and task completion, in real time through the Internet of Things module at a frequency of at least 10Hz, and then uses data processing and analysis technology to format and parse these data, and updates them to the user interface at a frequency of no less than 5Hz, ensuring that the operator can monitor the robot's dynamics in real time. The module also has a log recording function, which automatically records all operations and events, and the log data is retained for at least 30 days to facilitate troubleshooting and performance analysis. The built-in anomaly detection algorithm can identify abnormal conditions in the operation of the robot in real time with an accuracy of no less than 95%, and immediately notify relevant personnel when an abnormality is detected through various methods such as interface prompts, sounds, emails, and text messages. In addition, the module supports users to interact through an intuitive user interface and allows remote access and control with authorization, so that operators can access the monitoring system through the Internet from any location to configure parameters or execute control commands to improve production efficiency and safety.

[0040] The simulation module uses 3D visualization technology to simulate different production scenarios and robot behaviors with millimeter-level accuracy, ensuring a high degree of consistency between the simulation environment and the actual production environment. The module can simulate no less than 100 scenarios, covering a variety of production processes from simple to complex, to comprehensively evaluate the applicability and robustness of the scheduling strategy. The simulation accuracy reaches the millimeter level, ensuring that the robot's actions and production processes are highly consistent with the actual situation, so that the simulation results are highly reliable. It can simulate at least 100 different production scenarios, including production line layout, task type, workload and emergency situations, to ensure that the scheduling strategy can perform well in a variety of situations. Accurately simulate the robot's movement path, operation accuracy and task execution time, as well as the collaboration and communication between robots, to ensure that the simulation results reflect the actual performance of the robot. The effectiveness of the scheduling strategy is evaluated through the simulation results, including key indicators such as production efficiency, task completion time and resource utilization, to provide a basis for the optimization of the scheduling strategy. Provide a real-time feedback mechanism, allowing operators to adjust the policy parameters during the simulation process and immediately see the adjustment effect, so as to quickly iterate and optimize the scheduling strategy. Provides a data interface that allows direct import of scheduling strategies from the intelligent scheduling module and import of real-time data from the Internet of Things module for simulation analysis.

[0041] The human-machine interaction module is carefully designed to provide the ultimate user operation experience. It supports interface switching in at least 3 mainstream languages, including but not limited to Chinese, English and Spanish, to accommodate operators with different language backgrounds. The module uses the latest voice recognition technology, which can achieve a recognition accuracy of no less than 98% in noisy industrial environments, allowing operators to quickly and accurately control the robot through voice commands. At the same time, it also integrates an advanced gesture recognition system that can recognize and respond to the operator's gesture commands, as well as a highly sensitive touch screen interface to ensure intuitive and responsive operation. The combination of these functions greatly improves the convenience and intuitiveness of operation, reduces the difficulty of operation, and improves production efficiency.

[0042] Modular design modules are a key technology to improve the adaptability and flexibility of production lines. It allows operators to quickly replace or upgrade key components or modules of the robot, such as the end effector, sensor or control unit, within 2 hours to adapt to different production tasks or technology updates. This modular design reduces the time required for equipment conversion and maintenance, improves production efficiency, and reduces downtime caused by equipment adjustments. In addition, modular design also supports rapid fault diagnosis and repair, because the module can be tested and repaired independently of the main system, thereby reducing the downtime of the production line.

[0043] The Autonomous Mobile Robot (AMRs) module is designed to improve logistics efficiency. It is capable of handling and transporting materials at a speed of no less than 2m / s in a factory environment, ensuring rapid material flow and reducing production delays. The module integrates advanced path planning algorithms that can optimize transportation routes in real time to avoid congestion and conflicts. At the same time, the obstacle avoidance algorithm can respond within 0.5 seconds, ensuring that the path is quickly adjusted when an obstacle is detected to avoid collisions and ensure the safety of personnel and equipment. The combination of these functions enables the AMRs module to improve logistics efficiency while ensuring the safety and reliability of the material handling process.

[0044] The cloud computing platform module provides the system with powerful data processing and storage capabilities. It supports data processing capabilities of no less than 1TB / s, and can process a large amount of real-time data from the IoT module to ensure real-time data analysis and decision support. At the same time, the storage capacity provided by the module is no less than 100TB, which provides the system with sufficient data storage space, supports the storage and analysis of historical data, and the training and optimization of machine learning models. In addition, the cloud computing platform also supports remote monitoring and control functions, allowing operators to access the system from any location through a secure network connection, monitor production status in real time, and remotely adjust production parameters to ensure efficient operation and scalability of the system.

[0045] Through the collaborative work of the above modules, the multi-robot intelligent scheduling system of the present invention can realize the efficient collaborative operation of robots, improve the operating efficiency and accuracy of the production line, and meet the needs of modern manufacturing industry for flexible and efficient production processes.

[0046] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A multi-robot intelligent scheduling system based on deep learning, characterized in that: include: A deep learning module is used to predict the robot's operating efficiency and accuracy through historical data training, where the training data includes at least 5,000 historical operation samples, each of which contains at least 10 characteristic parameters; an Internet of Things module is used to collect robot status information in real time, including position, speed, and task completion status, with a collection frequency of no less than 10 Hz; an intelligent scheduling module is used to dynamically adjust the robot's task allocation based on the deep learning prediction results and real-time data, using a genetic algorithm or a particle swarm optimization algorithm for task allocation, where the algorithm parameters include a population size of no less than 50 and an iteration number of no less than 100; Real-time monitoring module, used to provide a real-time monitoring interface for the robot status. The monitoring interface is updated at a frequency of no less than 5Hz, and supports operators to adjust scheduling strategies based on monitoring information. It has a log recording function, and the log retention time is no less than 30 days; The simulation module is used to verify the feasibility of the scheduling strategy and supports 3D visualization to simulate the robot's operating environment and the effect of the scheduling strategy, with a simulation accuracy of millimeter level. The human-computer interaction module is used to provide a user-friendly operating interface, allowing operators to communicate and control the robot more easily, and supports at least 3 language interface switching. The modular design module allows for rapid replacement of robot parts or modules, with a replacement time of no more than 2 hours to adapt to different production tasks and improve the flexibility of the production line. The autonomous mobile robot (AMRs) module is used to perform material handling and transportation tasks, improve logistics efficiency, and the transportation speed is not less than 2m / s. The cloud computing platform module is used to realize data storage, processing and analysis, and provide remote monitoring and control functions. The data processing capacity is not less than 1TB / s and the storage capacity is not less than 100TB. The data preprocessing unit is used to clean and format the collected data to provide high-quality input data for the deep learning model, and the data cleaning accuracy is not less than 99%; A feature extraction unit is used to extract key features from the data to improve the prediction accuracy of the deep learning model, and the feature extraction accuracy is not less than 95%; a model training unit is used to train the deep learning model using the extracted features and labeled data, and the model training accuracy is not less than 90%; The task optimization unit is used to optimize the task allocation strategy and improve production efficiency based on the prediction results and real-time data of the deep learning model. The production efficiency after optimization is increased by no less than 20%; The communication management unit is used to manage and coordinate the communication between robots to ensure the accurate transmission of information and the smooth execution of tasks. The communication delay does not exceed 50ms. Safety monitoring unit, used to monitor the operating status of the robot to ensure the safety of the production process, with a safety monitoring coverage rate of 100%; The maintenance management unit is used to track the maintenance cycle and maintenance requirements of the robot to ensure the normal operation and extend the service life of the robot, and the maintenance prediction accuracy is not less than 90%. The Internet of Things module collects the status information of the robot in real time at a frequency of at least 10Hz, and transmits the data to the data preprocessing unit for cleaning and formatting to ensure that the deep learning model receives high-quality input data. Subsequently, the feature extraction unit extracts key features from the processed data, and the model training unit uses these features to train the deep learning model to predict the operating efficiency and accuracy of the robot. The intelligent scheduling module dynamically adjusts the robot task allocation using a genetic algorithm or a particle swarm optimization algorithm based on the prediction results of the deep learning model and real-time data. At the same time, the real-time monitoring module updates the monitoring interface of the robot status at a frequency of not less than 5Hz, so that the operator can adjust the scheduling strategy through the human-computer interaction module according to the monitoring information. The simulation module verifies the feasibility of the scheduling strategy in this process and displays the simulation effect through 3D visualization. The modular design module allows for rapid replacement of robot components or modules to adapt to different production tasks. The autonomous mobile robots (AMRs) module performs material handling and transportation tasks, while the cloud computing platform module is responsible for data storage, processing and analysis, and provides remote monitoring and control functions. The communication management unit ensures that the communication between robots is accurate and without delay, the safety monitoring unit is responsible for monitoring the operating status of the robots, and the maintenance management unit tracks the maintenance cycle and needs of the robots to ensure the stable operation of the system. The entire process forms a highly automated and intelligent closed-loop system, realizing full-process optimization from data collection, processing, analysis to decision-making, thereby significantly improving the efficiency and accuracy of multi-robot collaborative operations.

2. The system according to claim 1, characterized in that The deep learning module can automatically update and optimize the algorithm model, with an update cycle of no more than 1 month, to adapt to changes in the production environment.

3. The system according to claim 1, characterized in that The IoT module also includes an edge computing function for performing data processing near the data source to reduce data transmission delays. The response time of edge computing does not exceed 10ms.

4. The system according to claim 1, characterized in that The intelligent scheduling module can handle multi-task and multi-priority jobs, realize efficient scheduling of complex production tasks, and the scheduling strategy update frequency is not less than 1 time / minute.

5. The system according to claim 1, characterized in that The real-time monitoring module also includes an anomaly detection function, which can identify and respond to abnormal situations in the production process in real time, and the anomaly detection accuracy rate is not less than 95%.

6. The system according to claim 1, characterized in that The simulation module can simulate different production scenarios and robot behaviors to evaluate the effectiveness of scheduling strategies, with the number of simulated scenarios being no less than 100.

7. The system according to claim 1, characterized in that The human-computer interaction module supports voice recognition, gesture recognition and touch screen operation, improving the convenience and intuitiveness of operation, and the voice recognition accuracy is not less than 98%.

8. The system according to claim 1, characterized in that The modular design module allows the robot to quickly switch between different production tasks, reducing production preparation time, and the switching time does not exceed 1 hour. The system according to claim 1, characterized in that The autonomous mobile robots (AMRs) module includes path planning and obstacle avoidance algorithms to improve the efficiency and safety of material handling, with an obstacle avoidance response time of no more than 0.5 seconds.

Images