Industrial robot cluster control method, system and web server

Abstract

The application discloses an industrial robot cluster control method, system and network server, the industrial robot cluster comprises several robots, the control method comprises: obtaining state data affecting the communication or working state of the robot to obtain an initial data set; performing feature extraction processing on the initial data set to extract key features reflecting various state conditions of the robot in the initial data set, and forming a feature data set based on the key features; processing the feature data set based on a prediction model of a neural network architecture to predict the risk degree of the robot related to communication interruption; and adjusting the control strategy for the robot based on the risk degree, the control strategy being used for controlling the communication state and working state of the robot. The control method can effectively improve the stability of the robot communication, and can also adjust the working state of the robot to make the energy consumption of the robot within a reasonable range, so that the energy-saving effect of the industrial robot cluster is improved.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and in particular to a method, system, and network server for controlling industrial robot swarms based on mobile communication. Background Technology

[0002] Underground mining operations are complex and challenging tasks, often conducted in extreme environments. This work encompasses multiple phases, including search, monitoring, preparation, processing, maintenance, drilling, loading, and transporting underground materials to surface stations. These operating environments frequently face harsh temperature conditions, high humidity, high pressure, and complex and unstable geological structures, posing significant challenges to the normal operation of equipment and the stability of communication networks.

[0003] To improve the safety, efficiency, and productivity of mining operations, robotics and IoT devices have been introduced into the mining industry. These robots can perform automated tasks in underground environments, such as environmental monitoring, equipment inspection, and material handling. However, the successful execution of these tasks relies on efficient communication networks to ensure reliable data transmission and timely responses. In underground tunnels, interconnected communication subnetworks are typically formed, such as environmental IoT devices, which usually only require intermittent communication.

[0004] While these network architectures offer the necessary coverage and data transmission capabilities, they present numerous challenges in complex underground environments. These technologies generally suffer from severe signal attenuation, limited transmission distance, and high communication latency, making it difficult to provide stable communication connections in long-distance tunnels or complex multi-node environments. Furthermore, these network architectures typically employ static energy-saving strategies, failing to adequately consider dynamic changes in real-time environmental conditions and equipment health, resulting in deficiencies in energy consumption optimization and communication efficiency. Summary of the Invention

[0005] The purpose of this invention is to provide an industrial robot cluster control method, system, and network server that can dynamically adjust the robot's communication and working status based on the robot's current state.

[0006] To achieve the above objectives, the present invention provides a method for controlling an industrial robot swarm based on mobile communication. The industrial robot swarm includes several self-moving robots connected to a mobile communication network. The control method includes:

[0007] Acquire state data that affects the robot's communication or working state to obtain an initial dataset;

[0008] The initial dataset is processed by the computing nodes in the mobile communication network to extract key features that reflect the various states of the robot in the initial dataset, and a feature dataset is formed based on the key features.

[0009] The feature dataset is processed based on a pre-trained prediction model with a neural network architecture to predict the robot's risk level related to communication interruption.

[0010] Based on the risk level, the control strategy for the robot is adjusted, and the control strategy is used to control the robot's communication state and working state.

[0011] Preferably, the status data includes one or more of the following four types of data: first data related to the robot's current geographical location, second data related to the robot's current communication status, third data related to the robot's current health status, and fourth data related to the environmental status of the robot's current environment.

[0012] Preferably, the control strategy includes a first strategy for adjusting the robot based on the state data of the individual robot; and multiple sets of the first strategy are preset with respect to different levels of risk.

[0013] Preferably, the first strategy includes a device layer strategy and a network layer strategy, wherein the device layer strategy is used to control the working state of the robot, and the second strategy is used to control the communication state of the robot.

[0014] Preferably, the device layer strategy is used to adjust the robot's execution tasks to change power consumption; the network layer strategy is used to adjust the signal transmission rate, frequency, path, and redundancy of the communication network to which the robot belongs.

[0015] Preferably, the control strategy includes a second strategy for adjusting the working state and communication state of each robot at a global level based on the state data of multiple robots in the industrial robot cluster, wherein the second strategy has a higher priority than the first strategy.

[0016] Preferably, the second strategy coordinates and schedules the computing resources, network bandwidth, and energy allocated to each robot in the industrial robot cluster.

[0017] The present invention also provides a mobile communication-based network server for managing and controlling an industrial robot cluster connected thereto. The industrial robot cluster includes several robots, and the network server controls the working and communication states of the robots based on the industrial robot cluster control method described above.

[0018] The present invention also provides an industrial robot swarm control system, which includes:

[0019] One or more processors;

[0020] Memory;

[0021] and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including instructions for performing the industrial robot cluster control method as described above.

[0022] The present invention also provides a computer-readable storage medium comprising a computer program that can be executed by a processor to perform the industrial robot swarm control method described above.

[0023] Compared with the prior art, the industrial robot cluster control method disclosed in the above technical solution of the present invention acquires the state data of the robot during the working process, obtains the feature dataset that affects the interruption of robot communication through accurate feature extraction, and then processes the feature dataset through a prediction model to obtain the current risk level. Based on the risk level, the control strategy is dynamically adjusted, thereby effectively improving the stability of robot communication. It can also keep the robot's energy consumption within a reasonable range by adjusting the robot's working state, thereby improving the energy-saving effect of the industrial robot cluster. Attached Figure Description

[0024] Figure 1 This is a flowchart of the industrial robot cluster control method in an embodiment of the present invention.

[0025] Figure 2 This is a flowchart illustrating the signaling execution process of the network server in an embodiment of the present invention. Detailed Implementation

[0026] To illustrate the technical content, structural features, objectives, and effects of the present invention in detail, the following description is provided in conjunction with the embodiments and accompanying drawings.

[0027] This embodiment discloses a method for controlling an industrial robot swarm based on mobile communication, which dynamically controls the communication and working states of the robots to ensure communication stability and energy efficiency. The robots in this embodiment can be used in underground mining, but are not limited thereto.

[0028] like Figure 1 The control method includes the following steps:

[0029] S1: Obtain state data that affects the robot's communication or working state to obtain an initial dataset.

[0030] S2: Based on the computing power nodes in the mobile communication network, feature extraction processing is performed on the initial dataset to extract key features reflecting the various states of the robot, and a feature dataset is formed based on these key features. In this embodiment, the computing power node is a centralized edge technology node located in a base station. This centralized edge technology node receives and processes the initial dataset sent by all robots in the industrial robot cluster. The centralized edge technology node performs feature extraction processing on the initial dataset using an intelligent model with learning capabilities. Specifically, the linear regression model Lasso (Least Absolute Shrinkage and Selection Operator) is used for feature selection to automatically select important features, i.e., key features.

[0031] Additionally, it's worth noting that after sending the initial dataset to the centralized edge computing node, the data in the initial dataset can be filtered. Data with timestamps that don't meet requirements due to communication delays, data with excessively low signal strength, and data missing important fields should be removed to ensure the accuracy of subsequent feature extraction. Then, the Z-Score method is used to select outliers. Points with an absolute Z value greater than a preset value are considered outliers and deleted to prevent abnormal data from interfering with the analysis results.

[0032] S3: The feature dataset is processed based on a pre-trained prediction model with a neural network architecture to predict the robot's risk level related to communication interruption. In this embodiment, the prediction model is preferably a random forest model, which is also set in the base station; however, it can also be set in other network elements within its core network.

[0033] S4: Based on the risk level, adjust the control strategy for the robot, the control strategy being used to control the robot's communication state and working state.

[0034] Based on the aforementioned industrial robot swarm control method, the robots first collect various data affecting communication or operational status. These data are then filtered and extracted to form a feature dataset. This process enhances data quality, accurately identifies factors impacting communication stability, and provides a reliable data foundation for subsequent risk prediction. The accuracy of the feature dataset improves the identification of communication interruption risks, helping to prevent potential communication disruptions.

[0035] Specifically, by processing the feature dataset using a pre-trained neural network-based prediction model, and through sufficient learning and training of the model, the risk level associated with communication interruptions can be predicted more accurately. This enables timely adjustments to communication-related control strategies before communication interruptions occur, including switching communication paths and increasing signal transmission strength, thereby reducing or avoiding actual interruptions and improving the stability of the entire communication network.

[0036] Secondly, when the predictive model anticipates different levels of communication disruption risk for the robot, it can dynamically adjust the robot's control strategies related to its working state. For example, in low-risk situations, the existing state can be maintained to avoid unnecessary energy consumption or reduce communication consumption. In medium- and high-risk situations, appropriate adjustments and emergency measures can be taken respectively to maximize energy efficiency while ensuring communication stability. For instance, in medium-risk situations, alternative communication paths and data compression strategies might be selected to save energy; in high-risk situations, some non-core tasks might be delayed or stopped, concentrating energy on maintaining communication for critical tasks. This risk-adjustment strategy makes energy management more scientific, avoiding the inefficiency and resource waste of traditional static strategies.

[0037] In summary, the above control methods have significantly improved both robot communication and energy management.

[0038] On the other hand, the status data includes one or more of the following four types of data: first data related to the robot's current geographical location, second data related to the robot's current communication status, third data related to the robot's current health status, and fourth data related to the environmental status of the robot's current environment.

[0039] Specifically, 1. The first data includes one or more of the following:

[0040] 1.1 The robot's absolute position, such as longitude, latitude, and altitude;

[0041] 1.2 Relative position: The distance and direction between the robot and the base station, other robots, or fixed reference points;

[0042] 1.3 Rate of change of position: The robot's speed and direction of movement, which helps determine the robot's trajectory and predict its future position;

[0043] 1.4 Area Identifier: The specific area identifier of the robot's current working area (such as a mine), such as the mine number, working face number, etc.

[0044] It should be noted that the robot's absolute position and area identification are primary data collected directly from the robot, while the relative position and position change rate are secondary data calculated based on the positions of all robots and the base station.

[0045] 2. The second set of data includes one or more of the following:

[0046] 2.1 Signal Strength: Includes Received Signal Strength Indication (RSSI) and Signal-to-Noise Ratio (SNR);

[0047] 2.2 Data transmission rate: including uplink and downlink data transmission rates;

[0048] 2.3 Communication delay: The delay in data transmission from the robot to the base station, usually measured in milliseconds;

[0049] 2.4 Packet loss rate: The proportion of data packets lost during data transmission;

[0050] 2.5 Connection Status: Whether the connection is currently established and the stability of the connection.

[0051] 3. The third data includes one or more of the following:

[0052] 2.1 Battery Power: Remaining battery power and estimated runtime of the battery currently used as the power source in the robot;

[0053] 2.2 Running Time: The cumulative working time of the robot since its last charge;

[0054] 2.3 Hardware Health Log: Includes status monitoring data of various hardware components in the robot, such as temperature, vibration, and wear level;

[0055] 2.4 Fault Logs: These include recent and historical error logs and fault records from the robot, used to identify potential hardware problems.

[0056] 4. The third data includes one or more of the following:

[0057] 4.1 Temperature: The temperature of the robot's current operating environment;

[0058] 4.2 Humidity: The humidity level of the robot's current operating environment;

[0059] 4.3 Air Pressure: The air pressure value of the robot's current working environment;

[0060] 4.4 Air quality: The composition and concentration of gases in the robot's current working environment, such as carbon dioxide, carbon monoxide, and other harmful gases;

[0061] 4.5 Light Intensity: The lighting conditions in the robot's current working environment.

[0062] The significance of acquiring initial data regarding the robot's geographical location lies in:

[0063] First, the robot's geographical location and its relative position to the base station directly affect the transmission quality of communication signals;

[0064] Secondly, by tracking the robot's position in real time, network resources can be dynamically allocated through control strategies. This ensures that robots farther from the base station or with weaker signals receive more bandwidth or lower latency communication support. It also helps the base station's centralized edge computing nodes predict robot movement paths and perform load balancing based on these paths.

[0065] Furthermore, some tasks may have higher requirements for the robot's position. Therefore, by adjusting the control strategy, the priority of tasks can be adjusted according to the robot's position, ensuring that robots in critical locations can execute high-priority tasks first. For robots in non-critical areas, the system can postpone the execution of their tasks. In addition, if a robot is located in an area far from the base station or in a place with poor signal, the system may choose to reduce the robot's workload and prioritize tasks with lower communication requirements to avoid task failure due to communication interruption.

[0066] Finally, the relative position of the robot to the base station affects its communication energy consumption. Robots farther from the base station require more energy to maintain signal transmission. By detecting the position, control strategies can be dynamically adjusted based on communication consumption in different areas to reduce energy consumption. For robots at long distances, it may be necessary to switch to a more energy-efficient communication mode, such as reducing the communication frequency or delaying non-critical tasks, to reduce energy consumption.

[0067] Based on the relationship between the robot's position and communication status, before transmitting the initial dataset to the centralized edge technology node in the base station, the data can be preprocessed by a lightweight edge computing node located in the robot to correlate the position data with the signal strength, so that the key features extracted by the centralized edge technology node are more accurate and effective.

[0068] On the other hand, the control strategy includes a first strategy for adjusting the robot based on the state data of each individual robot. To this end, multiple sets of the first strategies are preset, each associated with a different level of risk. Specifically, when the risk level includes low, medium, and high risk, three sets of first strategies corresponding to low, medium, and high risk are pre-generated. Thus, when the prediction model outputs the risk level, a suitable first strategy can be selected based on that risk level and sent to the robot.

[0069] Furthermore, the first strategy includes a device layer strategy and a network layer strategy, wherein the device layer strategy is used to control the working state of the robot, and the second strategy is used to control the communication state of the robot.

[0070] Specifically, the device-layer strategy is used to adjust the robot's execution tasks to change power consumption. The network-layer strategy is used to adjust the signal transmission rate, frequency, path, and redundancy of the communication network to which the robot belongs.

[0071] For example, when the risk level (<30%) is low, it indicates that the current communication status, environmental conditions, and robot health are good. The primary objective then is to maintain the stability of the existing state and maximize energy efficiency. In this case, the control strategy mainly focuses on robot energy saving and communication resource optimization. The specific content of the device-level and network-level strategies is as follows:

[0072] Device-level strategy: Choose methods to reduce robot power consumption, such as lowering the sampling frequency of data sensors and reducing non-critical tasks of the robot. For example, for temperature monitoring sensors, the sampling frequency can be reduced from once per minute to once every 5 minutes. This can save power without affecting robot health and monitoring effectiveness.

[0073] Network layer strategy: Maintain the current communication path and frequency, but appropriately reduce the frequency of data transmission, especially in relatively stable environments where data changes are minimal. For example, if the distance between robots is fixed and they are not moving, the frequency of location data uploads can be reduced to once every 10 minutes or longer.

[0074] When the risk level (30%–70%) is at a medium risk level, it indicates that communication and robot health may be threatened to some extent, but not to the point of emergency. At this time, the control strategy needs to be adjusted to prevent risk escalation. The control strategy at this stage mainly focuses on proactive prevention and optimization of current task execution. The specific content of the device-level and network-level strategies is as follows:

[0075] Equipment-level strategy: Conduct more self-checks on the robot's health status and prioritize maintenance of its core tasks. This can reduce or postpone some secondary tasks, ensuring the robot focuses its resources on critical tasks. For example, a mining robot might temporarily suspend unimportant inspection tasks to concentrate on current material handling. Simultaneously, increase the frequency of battery status monitoring to ensure a stable power supply.

[0076] Network layer strategies: Communication paths can be adjusted in advance to prevent potential communication interruptions. For example, switching to a backup communication path or temporarily increasing the data transmission frequency can ensure timely transmission of critical data. If the current signal strength begins to weaken, repeaters or signal amplifiers can be activated in advance to extend the communication distance.

[0077] When the risk level (>70%) is high, immediate emergency measures are required to prevent major communication or equipment failures. The strategy at this time focuses on emergency handling and restoring the stability of communication or robot operation. The specific content of the device-level and network-level strategies is as follows:

[0078] Device-level strategy: The robot immediately enters a low-power mode, ceasing all non-critical tasks. For example, the robot will pause most sensor data acquisition, retaining only core health monitoring and self-check functions. The robot may reduce its operating frequency and minimize power-intensive tasks such as material handling to ensure the battery can last longer.

[0079] Network layer strategy: In high-risk situations, communication paths may be severely disrupted or about to be interrupted. In this case, the system should switch to the most stable communication path or use redundant communication methods, such as relay stations or additional signal amplifiers, to enhance current communication quality. Simultaneously, the data transmission frequency may be increased to ensure that as much important data as possible is transmitted during the final period of stable communication.

[0080] On the other hand, the control strategy includes a second strategy for adjusting the working state and communication state of each robot at a global level based on the state data of multiple robots in the industrial robot cluster, wherein the second strategy has a higher priority than the first strategy.

[0081] Specifically, the second strategy coordinates and schedules the computing resources, network bandwidth, and energy allocated to each robot in the industrial robot cluster.

[0082] In this embodiment, the second strategy is a system-level strategy. Through global resource management, it ensures that the overall energy consumption and performance of the entire industrial robot cluster can be maintained at a reasonable level under high load conditions. For example, at certain times, the number of robots in the industrial robot cluster surges or encounters harsh communication scenarios, and computing resources or bandwidth may face bottlenecks. At this time, the global resource management strategy prioritizes the allocation of resources to links with high communication load, ensuring uninterrupted communication and reducing energy consumption.

[0083] The second strategy, based on the system level, does not focus on solving the problem of a single robot, but rather considers the service quality and resource allocation of the entire industrial robot cluster. Under high load conditions (such as multiple robots simultaneously transmitting large amounts of data to the base station), a QoS adjustment strategy can be used to temporarily reduce the communication quality of low-priority tasks, thereby ensuring the communication stability of high-priority tasks (such as emergency fault alarms).

[0084] More specifically, the second strategy determines which tasks the robot should prioritize and which can be postponed, ensuring that critical tasks are completed with limited resources. When a mining robot encounters communication failures or equipment malfunctions in harsh environments, this second strategy prioritizes resource allocation to handle these urgent tasks, ensuring the robot can quickly return to normal operation. Non-urgent data acquisition tasks can be delayed until resources become available.

[0085] In another preferred embodiment of the present invention, a mobile communication-based network server is also disclosed, characterized in that the network server is used to manage and control an industrial robot cluster connected thereto, the industrial robot cluster including a number of robots, and the network server controls the working state and communication state of the robots based on the industrial robot cluster control method in the above embodiment.

[0086] Specifically, such as Figure 2 The robot sends the collected initial dataset to the base station, where the computing nodes are set up to perform feature extraction and risk prediction to obtain the risk level. The risk level is then transmitted to the Access and Mobility Management Function (AMF) in the 5G core network or the next-generation evolved core network. The AMF then forwards the risk level to the Policy Control Function (PCF). The PCF selects the corresponding control policy based on the received risk level and sends the control policy to the robot.

[0087] This invention also discloses an industrial robot swarm control system, which includes one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors. The programs include instructions for performing the industrial robot swarm control method as described above. The processors may be general-purpose central processing units (CPUs), microprocessors, application-specific integrated circuits (ASICs), or one or more integrated circuits, used to execute relevant programs to achieve the functions required by the modules in the industrial robot swarm control system of this application embodiment, or to execute the industrial robot swarm control method of this application method embodiment.

[0088] This invention also discloses a computer-readable storage medium comprising a computer program executable by a processor to perform the industrial robot swarm control method described above. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center integrating one or more available media. The available medium can be read-only memory (ROM), random access memory (RAM), or magnetic media, such as floppy disks, hard disks, magnetic tapes, magnetic disks, or optical media, such as digital versatile discs (DVDs), or semiconductor media, such as solid-state disks (SSDs).

[0089] This application also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the aforementioned industrial robot swarm control method.

[0090] The above-disclosed embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for controlling an industrial robot swarm based on mobile communication, characterized in that, The industrial robot cluster includes several self-moving robots connected to a mobile communication network, and the control method includes: Acquire state data that affects the robot's communication or working state to obtain an initial dataset; The initial dataset is processed by the computing nodes in the mobile communication network to extract key features that reflect the various states of the robot in the initial dataset, and a feature dataset is formed based on the key features. The feature dataset is processed based on a pre-trained prediction model with a neural network architecture to predict the robot's risk level related to communication interruption. Based on the risk level, the control strategy for the robot is adjusted, and the control strategy is used to control the communication state and working state of the robot; The control strategy includes a first strategy for adjusting the robot based on the state data of the individual robot; and multiple sets of the first strategy are preset with respect to different levels of risk. The first strategy includes a device layer strategy and a network layer strategy. The device layer strategy is used to control the working state of the robot, and the network layer strategy is used to control the communication state of the robot. The device layer strategy is used to adjust the robot's execution tasks to change power consumption; the network layer strategy is used to adjust the signal transmission rate, frequency, path, and redundancy of the communication network to which the robot belongs.

2. The industrial robot swarm control method based on mobile communication according to claim 1, characterized in that, The status data includes one or more of the following four types of data: first data related to the robot's current geographical location, second data related to the robot's current communication status, third data related to the robot's current health status, and fourth data related to the environmental status of the robot's current environment.

3. The industrial robot swarm control method based on mobile communication according to claim 1, characterized in that, The control strategy includes a second strategy for adjusting the working state and communication state of each robot at a global level based on the state data of multiple robots in the industrial robot cluster. The second strategy has a higher priority than the first strategy.

4. The industrial robot swarm control method based on mobile communication according to claim 3, characterized in that, The second strategy coordinates and schedules the computing resources, network bandwidth, and energy allocated to each robot in the industrial robot cluster.

5. A network server based on mobile communication, characterized in that, The network server is used to manage and control the industrial robot cluster connected therein, the industrial robot cluster including several robots, and the network server controls the working status and communication status of the robots based on the industrial robot cluster control method according to any one of claims 1 to 4.

6. An industrial robot swarm control system, characterized in that, include: One or more processors; Memory; And one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs including instructions for performing the industrial robot swarm control method as claimed in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, Includes a computer program that can be executed by a processor to perform the industrial robot cluster control method as described in any one of claims 1 to 4.

Images