Inspection robot charging control system and charging device thereof

Abstract

The application discloses a kind of inspection robot charging control system and charging device thereof, it is related to inspection robot technical field, and the optimal balance between the charging demand and task execution of inspection robot is realized by dynamic task and charging scheduling algorithm: arranging internet of things sensor in the inspection area, the position, electric quantity, current task state and environmental condition etc.data of robot are monitored in real time;And edge computing node is deployed in the inspection area, sensor data is handled and analyzed in real time, to reduce the burden of central server, reduce data transmission delay;Based on the characteristics of inspection task and the activity trajectory of robot, charging point is flexibly arranged in the inspection area;Through internet of things data analysis, the position and quantity of charging point are dynamically adjusted, to avoid the return problem caused by fixed charging point;Mobile charging site goes to the position of robot according to the position and electric quantity state of robot and charges actively, to reduce the return time of robot.

Description

Technical Field

[0001] This invention relates to the field of inspection robot technology, specifically to an inspection robot charging control system and its charging device. Background Technology

[0002] With the development of technologies such as computer vision and machine learning, the performance and functions of inspection robots have been continuously improved, and their applications are becoming more and more widespread, with market demand also growing.

[0003] For example, patent CN115986864A discloses a charging control system and method for a substation inspection robot. This system includes a battery management module and a charging management module that communicate with each other. The charging management module, when a valid limit signal from the charging device is detected, acquires information on the battery pack charging voltage, battery pack configuration, and battery cell materials, and controls the charger to output a matching charging voltage. The battery management module contains at least two independent chargers, both initially in a de-energized state. A corresponding charging strategy is generated based on the robot's battery pack voltage: when the robot's battery pack voltage is below a preset threshold, all chargers operate simultaneously; when the robot's battery pack voltage is not below the preset threshold, the chargers are used alternately in a time-sharing manner.

[0004] Current charging control systems for inspection robots have issues with charging station utilization. The efficiency of charging station utilization cannot be coordinated with the efficiency of robot task completion, resulting in charging station congestion and inspection task obstruction. When multiple robots need to charge simultaneously, charging station congestion can easily occur, causing robots to queue and affecting overall inspection efficiency. In addition, robots need to return to fixed charging stations for charging, leading to interruptions in inspection tasks and increased return travel time. Therefore, there is an urgent need for a charging control system and charging device for inspection robots to solve these problems. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a charging control system and charging device for inspection robots, solving the problems in existing technologies where charging stations are prone to congestion when multiple robots need to charge simultaneously, causing robots to queue and affecting overall inspection efficiency; and where robots need to return to fixed charging stations for charging, leading to interruptions in inspection tasks and increased return travel time.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention provides a charging control system for an inspection robot. The system architecture comprises a robot unit, an IoT sensor network, edge computing nodes, and a cloud platform. Edge computing nodes are deployed within the inspection area to construct the IoT sensor network, enabling real-time processing and analysis of sensor data. Based on the IoT sensor network and edge computing nodes, a multi-robot collaborative scheduling algorithm is employed, combining the robot's battery level, task urgency, current task progress, and charging station availability to generate global scheduling decisions. Simultaneously, different priorities are assigned based on task importance and urgency, ensuring uninterrupted charging for high-priority tasks. The overall architecture of the inspection robot charging control system, including the robot unit, IoT sensor network, edge computing nodes, and cloud platform, achieves global scheduling of robot charging and tasks. Through the multi-robot collaborative scheduling algorithm, an optimal balance between charging needs and task execution is achieved, improving inspection efficiency and reducing charging waiting time and task interruptions.

[0008] The present invention is further configured such that the robot unit includes: a power management module, a task management module, and a communication module;

[0009] IoT sensor networks include: a data acquisition module and a data transmission module;

[0010] Edge computing nodes include: a data processing module and a preliminary decision-making module;

[0011] The cloud platform includes: a data storage module, a data analysis module, and an advanced decision-making module;

[0012] The power management module includes:

[0013] The power monitoring unit monitors the robot's battery level in real time, sends power status data, and uses AI models to predict power consumption trends and plan charging in advance.

[0014] The battery health unit monitors the battery's temperature, voltage, and charge / discharge cycle count to predict battery life.

[0015] The task management module includes:

[0016] The task execution unit manages the tasks currently being executed by the robot, updates task progress in real time, and adjusts task priority and path planning in real time by combining edge computing.

[0017] The task scheduling unit collaborates with the cloud platform and edge computing nodes to dynamically adjust task allocation and execution order;

[0018] The communication module includes:

[0019] The wireless communication unit communicates with edge computing nodes and cloud platforms to transmit data and receive commands in real time.

[0020] The local communication unit communicates with nearby mobile charging stations to coordinate charging arrangements.

[0021] The data acquisition module includes:

[0022] The environmental monitoring unit collects environmental data, including temperature, humidity, and obstacle locations, to assist the robot in making decisions.

[0023] Location tracking unit for real-time positioning of robots and charging stations;

[0024] The data transmission module includes:

[0025] Wireless transmission unit: transmits data from sensors to edge computing nodes in real time via wireless communication protocols;

[0026] Data synchronization unit: synchronizes sensor data with cloud platform data;

[0027] The data processing module includes:

[0028] The real-time data processing unit processes the real-time data transmitted from the sensors and robot units, performing preliminary analysis and filtering.

[0029] The predictive model unit uses machine learning algorithms to predict the robot's power consumption and task completion time.

[0030] The preliminary decision-making module includes:

[0031] The task optimization unit adjusts task priorities and optimizes paths based on real-time data and prediction results;

[0032] The charging scheduling unit generates a preliminary charging scheduling plan based on power forecast and task urgency, and sends it to the cloud platform for confirmation.

[0033] Data storage module, including:

[0034] Historical data storage unit, used to store historical data of robot operation, for model training and trend analysis;

[0035] Real-time data storage unit, used to store real-time data transmitted from edge computing nodes;

[0036] The data analysis module includes:

[0037] Advanced analytics unit for in-depth data analysis;

[0038] The model training unit optimizes the prediction and scheduling model based on historical and real-time data;

[0039] Advanced decision-making module, including:

[0040] The global optimization unit generates a globally optimized task and charging schedule based on the states and tasks of all robots.

[0041] The present invention is further configured to construct an Internet of Things (IoT) sensor network, deploying multiple sub-regions within the inspection area, with multiple robots and sensors distributed within each sub-region, and deploying edge computing nodes along the main inspection path; by deploying sensors and edge computing nodes in different regions, data transmission latency is reduced, and the real-time performance and accuracy of data processing are improved.

[0042] The entire inspection area is divided into multiple sub-areas, and multiple robots and multiple sensors are distributed in each sub-area.

[0043] Edge computing nodes are deployed along the main inspection paths and in areas with frequent tasks within the sub-region. Sensors and robots in each sub-region communicate with the neighboring edge computing nodes and send data to the neighboring edge computing nodes.

[0044] The node receives and caches real-time data streams, performs data cleaning and preprocessing, filters noisy data, and formats the raw data.

[0045] Conduct data analysis and decision-making, including preliminary analysis of data, such as power generation forecasting and task progress assessment;

[0046] Based on the analysis results, preliminary task adjustment and charging scheduling decisions are generated and sent to the robot unit.

[0047] Regularly upload the processed analysis results and model updates to the cloud platform;

[0048] Regularly retrieve the latest machine learning models from the cloud platform and update the models on the edge computing nodes;

[0049] The present invention is further configured to perform data cleaning and preprocessing, use a moving average method for data smoothing, and use LSTM for power prediction, thereby improving the accuracy of power prediction and task progress assessment and optimizing scheduling decisions.

[0050] Data cleaning and preprocessing were performed to remove erroneous sensor readings; the raw data was converted to a uniform format and normalized. scope: ,in It is the raw data. It is normalized data. It is the minimum value of the data. It is the maximum value of the data;

[0051] Use the moving average method to smooth the data: ,in For a moment Smoothed data, For a moment The original data, It's a smooth window size. Indicates the index of the data points within the smoothing window;

[0052] Energy prediction using a Long Short-Term Memory (LSTM) network, model formula:

[0053] ;

[0054] ;

[0055] ;

[0056] ,in For the hidden state vector, This represents the input vector, including the current battery level and task status. It is the Sigmoid activation function. , This is the weight matrix. and Indicates the bias term. Indicates the state of the memory cell. It is the gating unit of LSTM;

[0057] The LSTM model is trained using historical electricity data, and the weights and biases are adjusted using the backpropagation algorithm.

[0058] Predict future battery levels using a trained LSTM model: ,in It is the predicted battery level for the next moment. Indicates the current battery level;

[0059] Perform task progress assessment; task status parameters include task start time. Expected completion time Current time Task progress The calculation formula is as follows: ,in Indicates the percentage of task progress. Indicates the task start time. Indicates the estimated completion time of the task. The current time;

[0060] Predicting task completion time using a linear regression model: ,in Indicates the predicted completion time. This represents the task feature vector, including task type, task difficulty, and current progress. Represents the weight vector. For bias terms;

[0061] Make preliminary task adjustments and charging scheduling decisions;

[0062] The present invention is further configured to achieve efficient scheduling of tasks and charging by optimizing the model and constraints, reducing robot waiting time, and ensuring the continuous execution of high-priority tasks. The preliminary task adjustment and charging scheduling decision-making steps include:

[0063] Construct a decision model, with decision parameters including robot power. Task Priority Task progress Charging station availability The optimization objective is defined as: minimizing task interruption time and charging wait time, while maximizing the continuity of high-priority tasks. Therefore: ,in These represent weighting coefficients, indicating the importance of task interruption time, charging wait time, and task priority, respectively. Indicates task Interruption time, It is a charging station The waiting time Indicates task Priority factor;

[0064] The constraints are:

[0065] Robot power constraints: ,in It's a robot. At any moment The amount of electricity, This is the minimum safe power level;

[0066] Charging station capacity constraints: ,in It is a charging station At any moment Availability, Indicates the maximum capacity of the charging station;

[0067] Based on power prediction and task progress assessment, robots with low power are prioritized for dispatch to available charging stations, and a preliminary dispatch plan is generated based on task priority. ,in It refers to the scheduling scheme;

[0068] The initial task adjustment and charging scheduling decisions are sent to the robot unit to execute the scheduling instructions.

[0069] The present invention is further configured such that the global scheduling decision output method is:

[0070] Assign different priorities (high, medium, low) based on the importance and urgency of the task, and define task priority parameters;

[0071] A global scheduling decision model is constructed, with the objective function being to minimize task interruption time and charging wait time while maximizing the continuity of high-priority tasks.

[0072] The constraints include robot power constraints, charging station capacity constraints, and task priority constraints.

[0073] Initial scheduling decisions are made on edge computing nodes based on real-time data, and robots are assigned to charging stations.

[0074] Edge computing nodes perform local optimization based on real-time data, adjusting tasks and charging schedules accordingly.

[0075] The initial decisions and data from edge computing nodes are uploaded to the cloud platform for comprehensive global analysis and optimization.

[0076] Reinforcement learning (RL) is used to perform global optimization on a cloud platform to generate scheduling schemes.

[0077] The globally optimized scheduling scheme is distributed to each robot through edge computing nodes to execute scheduling instructions;

[0078] Edge computing nodes and cloud platforms monitor execution in real time and dynamically adjust scheduling schemes to deal with emergencies; through global scheduling decisions and real-time optimization, the system's response speed and overall efficiency are improved.

[0079] The present invention is further configured such that the global scheduling decision output step includes:

[0080] Each task is assigned a priority parameter. The priority parameter is divided into three values: high, medium, and low. ,middle ,Low ;

[0081] A global scheduling decision model is constructed, defining the objective function as: minimizing task interruption time and charging wait time, while maximizing the continuity of high-priority tasks, i.e. The constraints also include robot power constraints, charging station capacity constraints, and task priority constraints.

[0082] On edge computing nodes, preliminary scheduling decisions are made based on real-time data, robots are assigned to charging stations, and optimizations are made based on current battery level, task progress, and charging station availability.

[0083] The initial decisions and data from edge computing nodes are uploaded to the cloud platform for comprehensive global analysis and optimization.

[0084] Global optimization is performed on a cloud platform using reinforcement learning (RL) algorithms. The state, action, reward, and objective function of the RL model are defined as follows:

[0085] state The current system status, including robot battery level, task status, and charging station availability;

[0086] action Scheduling decisions under the current state, including task allocation and charging plans;

[0087] award Immediate reward calculated based on the objective function: ;

[0088] Maximize cumulative reward : ,in Discount factor;

[0089] The scheduling scheme optimized by RL on the cloud platform is distributed to each robot through edge computing nodes to execute scheduling instructions;

[0090] The edge computing nodes and cloud platform monitor the execution status in real time and dynamically adjust the scheduling plan to deal with emergencies.

[0091] This invention also provides a charging device for a patrol robot charging control system. The charging device utilizes the aforementioned patrol robot charging control system. Based on the characteristics of the patrol task and the robot's movement trajectory, the charging device dynamically adjusts the position and number of charging points based on the IoT data analysis results from the data analysis module. The charging device includes:

[0092] Fixed charging stations include:

[0093] The charging management unit monitors the charging process;

[0094] The communication unit communicates with the robot unit and the cloud platform to receive and execute charging scheduling instructions.

[0095] Mobile charging stations, including:

[0096] The positioning and navigation unit actively moves to the vicinity of the robot to recharge based on the robot's location and charging needs;

[0097] The charging coordination unit coordinates the charging time and sequence with multiple robots.

[0098] The present invention is further configured such that the charging device is arranged as follows:

[0099] Collect historical inspection task data, analyze the frequency, location and path of inspection tasks, and output key areas and high-frequency inspection paths of inspection tasks.

[0100] Based on the task analysis results, initial fixed charging points are set up in the inspection area, covering areas with frequent tasks and path intersections; reducing the robot's initial return distance.

[0101] IoT sensors are deployed in the inspection area to monitor the robot's location and battery level in real time, and upload the data to edge computing nodes and cloud platforms in real time.

[0102] The location and number of charging points are dynamically adjusted as follows:

[0103] Real-time monitoring of the usage of each fixed charging point, and statistics on charging frequency, charging duration and congestion;

[0104] Based on real-time and historical data, the location and number of charging points are dynamically adjusted; temporary charging points are added in areas with high charging demand.

[0105] The mobile charging station is equipped with a navigation system that actively heads to the robot's location to charge based on the robot's position and battery status.

[0106] Compared with the prior art, the present invention has the following beneficial effects:

[0107] This application, based on IoT technology and edge computing, achieves an optimal balance between charging needs and task execution for inspection robots through dynamic task and charging scheduling algorithms:

[0108] This invention deploys IoT sensors within the inspection area to monitor data such as the robot's location, battery level, current task status, and environmental conditions in real time; and deploys edge computing nodes within the inspection area to process and analyze sensor data in real time, reducing the burden on the central server and minimizing data transmission latency.

[0109] This invention, based on the characteristics of inspection tasks and the robot's activity trajectory, flexibly deploys charging points within the inspection area; through IoT data analysis, it dynamically adjusts the location and number of charging points to avoid the return trip problem caused by fixed charging points; mobile charging stations actively go to the robot's location to charge based on the robot's location and battery status, reducing the robot's return trip time.

[0110] This invention, based on IoT and edge computing technologies, employs a multi-robot collaborative scheduling algorithm. It combines the battery level of each robot, the urgency of the task, the current task progress, and the availability of charging stations to make global scheduling decisions. At the same time, it assigns different priorities according to the importance and urgency of the task, ensuring that high-priority tasks are not interrupted while the robot is charging.

[0111] This invention solves the problems in existing technologies, such as congestion at charging stations when multiple robots need to charge simultaneously, leading to robots queuing and affecting overall inspection efficiency; and the need for robots to return to fixed charging stations for charging, resulting in interruptions to inspection tasks and increased return travel time. Attached Figure Description

[0112] Figure 1 This is a framework diagram of the charging control system for the inspection robot of the present invention. Detailed Implementation

[0113] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0114] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0115] The present invention will now be described in further detail with reference to the accompanying drawings:

[0116] Example 1

[0117] Please see Figure 1This invention provides a charging control system for an inspection robot. The architecture of the inspection robot charging control system includes: a robot unit, an Internet of Things (IoT) sensor network, edge computing nodes, and a cloud platform. Edge computing nodes are deployed within the inspection area to construct the IoT sensor network and process and analyze sensor data in real time. Based on the IoT sensor network and edge computing nodes, a multi-robot collaborative scheduling algorithm is adopted, which combines the battery level of each robot, the urgency of the task, the current task progress, and the availability of charging stations to make global scheduling decisions. At the same time, different priorities are assigned according to the importance and urgency of the task to ensure that the charging of robots with high-priority tasks is not interrupted.

[0118] The robot unit includes: a power management module, a task management module, and a communication module;

[0119] IoT sensor networks include: a data acquisition module and a data transmission module;

[0120] Edge computing nodes include: a data processing module and a preliminary decision-making module;

[0121] The cloud platform includes: a data storage module, a data analysis module, and an advanced decision-making module;

[0122] The power management module includes:

[0123] The power monitoring unit monitors the robot's battery level in real time, sends power status data, and uses AI models to predict power consumption trends and plan charging in advance.

[0124] The battery health unit monitors the battery's temperature, voltage, and charge / discharge cycle count to predict battery life.

[0125] The task management module includes:

[0126] The task execution unit manages the tasks currently being executed by the robot, updates task progress in real time, and adjusts task priority and path planning in real time by combining edge computing.

[0127] The task scheduling unit collaborates with the cloud platform and edge computing nodes to dynamically adjust task allocation and execution order;

[0128] The communication module includes:

[0129] The wireless communication unit communicates with edge computing nodes and cloud platforms to transmit data and receive commands in real time.

[0130] The local communication unit communicates with nearby mobile charging stations to coordinate charging arrangements.

[0131] The data acquisition module includes:

[0132] The environmental monitoring unit collects environmental data, including temperature, humidity, and obstacle locations, to assist the robot in making decisions.

[0133] Location tracking unit for real-time positioning of robots and charging stations;

[0134] The data transmission module includes:

[0135] Wireless transmission unit: transmits data from sensors to edge computing nodes in real time via wireless communication protocols;

[0136] Data synchronization unit: synchronizes sensor data with cloud platform data;

[0137] The data processing module includes:

[0138] The real-time data processing unit processes the real-time data transmitted from the sensors and robot units, performing preliminary analysis and filtering.

[0139] The predictive model unit uses machine learning algorithms to predict the robot's power consumption and task completion time.

[0140] The preliminary decision-making module includes:

[0141] The task optimization unit adjusts task priorities and optimizes paths based on real-time data and prediction results;

[0142] The charging scheduling unit generates a preliminary charging scheduling plan based on power forecast and task urgency, and sends it to the cloud platform for confirmation.

[0143] Data storage module, including:

[0144] Historical data storage unit, used to store historical data of robot operation, for model training and trend analysis;

[0145] Real-time data storage unit, used to store real-time data transmitted from edge computing nodes;

[0146] The data analysis module includes:

[0147] Advanced analytics unit for in-depth data analysis;

[0148] The model training unit optimizes the prediction and scheduling model based on historical and real-time data;

[0149] Advanced decision-making module, including:

[0150] The global optimization unit generates a globally optimized task and charging schedule based on the states and tasks of all robots.

[0151] The method for building an IoT sensor network and processing and analyzing sensor data in real time is as follows:

[0152] The entire inspection area is divided into multiple sub-areas, and multiple robots and multiple sensors are distributed in each sub-area.

[0153] Edge computing nodes are deployed along the main inspection paths and in areas with frequent tasks within the sub-region. Sensors and robots in each sub-region communicate with the neighboring edge computing nodes and send data to the neighboring edge computing nodes.

[0154] The node receives and caches real-time data streams, performs data cleaning and preprocessing, filters noisy data, and formats the raw data.

[0155] Conduct data analysis and decision-making, including preliminary analysis of data, such as power generation forecasting and task progress assessment;

[0156] Based on the analysis results, preliminary task adjustment and charging scheduling decisions are generated and sent to the robot unit.

[0157] Regularly upload the processed analysis results and model updates to the cloud platform;

[0158] Regularly retrieve the latest machine learning models from the cloud platform and update the models on the edge computing nodes;

[0159] The data analysis and decision-making methods are as follows:

[0160] Data cleaning and preprocessing were performed to remove erroneous sensor readings; the raw data was converted to a uniform format and normalized. scope: ,in It is the raw data. It is normalized data. It is the minimum value of the data. It is the maximum value of the data;

[0161] Use the moving average method to smooth the data: ,in For a moment Smoothed data, For a moment The original data, It's a smooth window size. Indicates the index of the data points within the smoothing window;

[0162] Energy prediction using a Long Short-Term Memory (LSTM) network, model formula:

[0163] ;

[0164] ;

[0165] ;

[0166] ,in For the hidden state vector, This represents the input vector, including the current battery level and task status. It is the Sigmoid activation function. , This is the weight matrix. and Indicates the bias term. Indicates the state of the memory cell. It is the gating unit of LSTM;

[0167] The LSTM model is trained using historical electricity data, and the weights and biases are adjusted using the backpropagation algorithm.

[0168] Predict future battery levels using a trained LSTM model: ,in It is the predicted battery level for the next moment. Indicates the current battery level;

[0169] Perform task progress assessment; task status parameters include task start time. Expected completion time Current time Task progress The calculation formula is as follows: ,in Indicates the percentage of task progress. Indicates the task start time. Indicates the estimated completion time of the task. The current time;

[0170] Predicting task completion time using a linear regression model: ,in Indicates the predicted completion time. This represents the task feature vector, including task type, task difficulty, and current progress. Represents the weight vector. For bias terms;

[0171] Make preliminary task adjustments and charging scheduling decisions;

[0172] The initial task adjustment and charging scheduling decision-making steps include:

[0173] Construct a decision model, with decision parameters including robot power. Task Priority Task progress Charging station availability The optimization objective is defined as: minimizing task interruption time and charging wait time, while maximizing the continuity of high-priority tasks. Therefore: ,in These represent weighting coefficients, indicating the importance of task interruption time, charging wait time, and task priority, respectively. Indicates task Interruption time, It is a charging station The waiting time Indicates task Priority factor;

[0174] The constraints are:

[0175] Robot power constraints: ,in It's a robot. At any moment The amount of electricity, This is the minimum safe power level;

[0176] Charging station capacity constraints: ,in It is a charging station At any moment Availability, Indicates the maximum capacity of the charging station;

[0177] Based on power prediction and task progress assessment, robots with low power are prioritized for dispatch to available charging stations, and a preliminary dispatch plan is generated based on task priority. ,in It refers to the scheduling scheme;

[0178] The initial task adjustment and charging scheduling decisions are sent to the robot unit to execute the scheduling instructions.

[0179] The global scheduling decision output method is as follows:

[0180] Assign different priorities (high, medium, low) based on the importance and urgency of the task, and define task priority parameters;

[0181] A global scheduling decision model is constructed, with the objective function being to minimize task interruption time and charging wait time while maximizing the continuity of high-priority tasks.

[0182] The constraints include robot power constraints, charging station capacity constraints, and task priority constraints.

[0183] Initial scheduling decisions are made on edge computing nodes based on real-time data, and robots are assigned to charging stations.

[0184] Edge computing nodes perform local optimization based on real-time data, adjusting tasks and charging schedules accordingly.

[0185] The initial decisions and data from edge computing nodes are uploaded to the cloud platform for comprehensive global analysis and optimization.

[0186] Reinforcement learning (RL) is used to perform global optimization on a cloud platform to generate scheduling schemes.

[0187] The globally optimized scheduling scheme is distributed to each robot through edge computing nodes to execute scheduling instructions;

[0188] The edge computing nodes and cloud platform monitor the execution status in real time and dynamically adjust the scheduling plan to deal with emergencies.

[0189] The global scheduling decision output steps include:

[0190] Each task is assigned a priority parameter. The priority parameter is divided into three values: high, medium, and low. ,middle ,Low ;

[0191] A global scheduling decision model is constructed, defining the objective function as: minimizing task interruption time and charging wait time, while maximizing the continuity of high-priority tasks, i.e. The constraints also include robot power constraints, charging station capacity constraints, and task priority constraints.

[0192] On edge computing nodes, preliminary scheduling decisions are made based on real-time data, robots are assigned to charging stations, and optimizations are made based on current battery level, task progress, and charging station availability.

[0193] The initial decisions and data from edge computing nodes are uploaded to the cloud platform for comprehensive global analysis and optimization.

[0194] Global optimization is performed on a cloud platform using reinforcement learning (RL) algorithms. The state, action, reward, and objective function of the RL model are defined as follows:

[0195] state The current system status, including robot battery level, task status, and charging station availability;

[0196] action Scheduling decisions under the current state, including task allocation and charging plans;

[0197] award Immediate reward calculated based on the objective function: ;

[0198] Maximize cumulative reward : ,in Discount factor;

[0199] The scheduling scheme optimized by RL on the cloud platform is distributed to each robot through edge computing nodes to execute scheduling instructions;

[0200] The edge computing nodes and cloud platform monitor the execution status in real time and dynamically adjust the scheduling plan to deal with emergencies.

[0201] This invention also provides a charging device for a charging control system of an inspection robot. The charging device dynamically adjusts the position and number of charging points based on the characteristics of the inspection task and the robot's movement trajectory, and based on the IoT data analysis results from the data analysis module. The charging device includes:

[0202] Fixed charging stations include:

[0203] The charging management unit monitors the charging process;

[0204] The communication unit communicates with the robot unit and the cloud platform to receive and execute charging scheduling instructions.

[0205] Mobile charging stations, including:

[0206] The positioning and navigation unit actively moves to the vicinity of the robot to recharge based on the robot's location and charging needs;

[0207] The charging coordination unit coordinates the charging time and sequence with multiple robots.

[0208] The charging device is arranged as follows:

[0209] Collect historical inspection task data, analyze the frequency, location and path of inspection tasks, and output key areas and high-frequency inspection paths of inspection tasks.

[0210] Based on the task analysis results, initial fixed charging points are set up in the inspection area, covering areas with frequent tasks and path intersections; reducing the robot's initial return distance.

[0211] IoT sensors are deployed in the inspection area to monitor the robot's location and battery level in real time, and upload the data to edge computing nodes and cloud platforms in real time.

[0212] The location and number of charging points are dynamically adjusted as follows:

[0213] Real-time monitoring of the usage of each fixed charging point, and statistics on charging frequency, charging duration and congestion;

[0214] Based on real-time and historical data, the location and number of charging points are dynamically adjusted; temporary charging points are added in areas with high charging demand.

[0215] The mobile charging station is equipped with a navigation system that actively heads to the robot's location to charge based on the robot's position and battery status.

[0216] This invention deploys edge computing nodes within the inspection area to construct an Internet of Things sensor network, enabling real-time processing and analysis of sensor data. The system employs a multi-robot collaborative scheduling algorithm, combining each robot's battery level, task urgency, current task progress, and charging station status to make global scheduling decisions, ensuring that robots performing high-priority tasks are not interrupted while charging.

[0217] The robot unit includes a power management module, a task management module, and a communication module. The power management module monitors the battery level in real time through a power monitoring unit, uses an AI model to predict power consumption trends, and plans charging in advance. The battery health unit monitors the battery's temperature, voltage, and charge / discharge cycle count to estimate battery life. The task management module includes a task execution unit that manages current tasks, updates task progress in real time, and adjusts task priorities and path planning in real time using edge computing. The task scheduling unit collaborates with the cloud platform and edge computing nodes to dynamically adjust task allocation and execution order. The communication module communicates with edge computing nodes and the cloud platform through a wireless communication unit for real-time data transmission and command reception, and communicates with nearby mobile charging stations through a local communication unit to coordinate charging arrangements.

[0218] The Internet of Things (IoT) sensor network includes a data acquisition module and a data transmission module. The data acquisition module consists of an environmental monitoring unit and a location tracking unit. The former collects environmental data to assist robot decision-making, while the latter locates the robot and charging stations in real time. The data transmission module transmits sensor data to edge computing nodes in real time through a wireless transmission unit and synchronizes data with the cloud platform through a data synchronization unit.

[0219] The edge computing node includes a data processing module and a preliminary decision-making module. The data processing module is responsible for processing real-time data transmitted from sensors and robot units, performing preliminary analysis and filtering, and using machine learning algorithms through the prediction model unit to predict robot power consumption and task completion time. The preliminary decision-making module, through the task optimization unit and the charging scheduling unit, adjusts task priorities and optimizes paths based on real-time data and prediction results, generates a preliminary charging scheduling plan, and sends it to the cloud platform for confirmation.

[0220] The cloud platform includes a data storage module, a data analysis module, and an advanced decision-making module. The data storage module stores historical and real-time data and supports model training and trend analysis. The data analysis module performs in-depth data analysis through the advanced analysis unit and optimizes prediction and scheduling models based on historical and real-time data through the model training unit. The advanced decision-making module generates globally optimized task and charging scheduling schemes based on the state and tasks of all robots through the global optimization unit.

[0221] The charging device is arranged based on the characteristics of the inspection task and the robot's activity trajectory, and dynamically adjusts the location and number of charging points based on the results of IoT data analysis. Fixed charging stations include a charging management unit and a communication unit, which are responsible for monitoring the charging process and communicating with the robot unit and the cloud platform to execute charging scheduling instructions. Mobile charging stations include a positioning and navigation unit and a charging coordination unit, which can actively move to the vicinity of the robot for charging according to the robot's location and charging needs, and coordinate the charging time and sequence with multiple robots.

[0222] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A patrol robot charging control system, the patrol robot charging control system architecture comprising: The system comprises a robot unit, an IoT sensor network, edge computing nodes, and a cloud platform. Its key feature is the deployment of edge computing nodes within the inspection area to construct an IoT sensor network, enabling real-time processing and analysis of sensor data. Based on the IoT sensor network and edge computing nodes, a multi-robot collaborative scheduling algorithm is employed, combining each robot's battery level, task urgency, current task progress, and charging station availability to generate a global scheduling decision. Simultaneously, different priorities are assigned based on task importance and urgency, ensuring uninterrupted charging for high-priority robots. The global scheduling decision output method is as follows: Assign different priorities based on the importance and urgency of the tasks, and define task priority parameters. A global scheduling decision model is constructed, with the objective function being to minimize task interruption time and charging wait time while maximizing the continuity of high-priority tasks. The constraints include robot power constraints, charging station capacity constraints, and task priority constraints. Initial scheduling decisions are made on edge computing nodes based on real-time data, and robots are assigned to charging stations. Edge computing nodes perform local optimization based on real-time data, adjusting tasks and charging schedules accordingly. The initial decisions and data from edge computing nodes are uploaded to the cloud platform for comprehensive global analysis and optimization. Reinforcement learning (RL) is used to perform global optimization on a cloud platform to generate scheduling schemes. The globally optimized scheduling scheme is distributed to each robot through edge computing nodes to execute scheduling instructions; Edge computing nodes and cloud platforms monitor execution in real time and dynamically adjust scheduling plans to deal with emergencies.

2. The inspection robot charging control system according to claim 1, characterized in that, The robot unit includes: a power management module, a task management module, and a communication module; IoT sensor networks include: a data acquisition module and a data transmission module; Edge computing nodes include: a data processing module and a preliminary decision-making module; The cloud platform includes: a data storage module, a data analysis module, and an advanced decision-making module; The power management module includes: The power monitoring unit monitors the robot's battery level in real time, sends power status data, and uses AI models to predict power consumption trends and plan charging in advance. The battery health unit monitors the battery's temperature, voltage, and charge / discharge cycle count to predict battery life. The task management module includes: The task execution unit manages the tasks currently being executed by the robot, updates task progress in real time, and adjusts task priority and path planning in real time by combining edge computing. The task scheduling unit collaborates with the cloud platform and edge computing nodes to dynamically adjust task allocation and execution order; The communication module includes: The wireless communication unit communicates with edge computing nodes and cloud platforms to transmit data and receive commands in real time. The local communication unit communicates with nearby mobile charging stations to coordinate charging arrangements. The data acquisition module includes: The environmental monitoring unit collects environmental data, including temperature, humidity, and obstacle locations, to assist the robot in making decisions. Location tracking unit for real-time positioning of robots and charging stations; The data transmission module includes: Wireless transmission unit: transmits data from sensors to edge computing nodes in real time via wireless communication protocols; Data synchronization unit: synchronizes sensor data with cloud platform data; The data processing module includes: The real-time data processing unit processes the real-time data transmitted from the sensors and robot units, performing preliminary analysis and filtering. The predictive model unit uses machine learning algorithms to predict the robot's power consumption and task completion time. The preliminary decision-making module includes: The task optimization unit adjusts task priorities and optimizes paths based on real-time data and prediction results; The charging scheduling unit generates a preliminary charging scheduling plan based on power forecast and task urgency, and sends it to the cloud platform for confirmation. Data storage module, including: Historical data storage unit, used to store historical data of robot operation, for model training and trend analysis; Real-time data storage unit, used to store real-time data transmitted from edge computing nodes; The data analysis module includes: Advanced analytics unit for in-depth data analysis; The model training unit optimizes the prediction and scheduling model based on historical and real-time data; Advanced decision-making module, including: The global optimization unit generates globally optimized task and charging scheduling schemes based on the states and tasks of all robots.

3. The inspection robot charging control system according to claim 2, characterized in that, The method for building an IoT sensor network and processing and analyzing sensor data in real time is as follows: The entire inspection area is divided into multiple sub-areas, and multiple robots and multiple sensors are distributed in each sub-area. Edge computing nodes are deployed along the main inspection paths and in areas with frequent tasks within the sub-region. Sensors and robots in each sub-region communicate with the neighboring edge computing nodes and send data to the neighboring edge computing nodes. The node receives and caches real-time data streams, performs data cleaning and preprocessing, filters noisy data, and formats the raw data. Conduct data analysis and decision-making, including preliminary analysis of data, such as power generation forecasting and task progress assessment; Based on the analysis results, preliminary task adjustment and charging scheduling decisions are generated and sent to the robot unit. Regularly upload the processed analysis results and model updates to the cloud platform; Regularly retrieve the latest machine learning models from the cloud platform and update the models on the edge computing nodes.

4. The inspection robot charging control system according to claim 3, characterized in that, The data analysis and decision-making methods are as follows: Data cleaning and preprocessing were performed to remove erroneous sensor readings; the raw data was converted to a uniform format and normalized. scope: ,in It is the raw data. It is normalized data. It is the minimum value of the data. It is the maximum value of the data; Use the moving average method to smooth the data: ,in For a moment Smoothed data, For a moment The original data, It's a smooth window size. Indicates the index of the data points within the smoothing window; Energy prediction using a Long Short-Term Memory (LSTM) network, model formula: ； ； ； ,in For the hidden state vector, This represents the input vector, including the current battery level and task status. It is the Sigmoid activation function. , This is the weight matrix. and Indicates the bias term. Indicates the state of the memory cell. It is the gating unit of LSTM; The LSTM model is trained using historical electricity data, and the weights and biases are adjusted using the backpropagation algorithm. Predict future battery levels using a trained LSTM model: ,in It is the predicted battery level for the next moment. Indicates the current battery level; Perform task progress assessment; task status parameters include task start time. Expected completion time Current time Task progress The calculation formula is as follows: ,in Indicates the percentage of task progress. Indicates the task start time. Indicates the estimated completion time of the task. The current time; Predicting task completion time using a linear regression model: ,in Indicates the predicted completion time. This represents the task feature vector, including task type, task difficulty, and current progress. Represents the weight vector. For bias terms; Make preliminary task adjustments and charging scheduling decisions.

5. A charging control system for an inspection robot according to claim 4, characterized in that, The initial task adjustment and charging scheduling decision-making steps include: Construct a decision model, with decision parameters including robot power. Task priority Task progress Charging station availability The optimization objective is defined as: minimizing task interruption time and charging wait time, while maximizing the continuity of high-priority tasks. Therefore: ,in These represent weighting coefficients, indicating the importance of task interruption time, charging wait time, and task priority, respectively. Indicates task Interruption time, It is a charging station The waiting time Indicates task Priority factor; The constraints are: Robot power constraints: ,in It's a robot. At any moment The amount of electricity, This is the minimum safe power level; Charging station capacity constraints: ,in It is a charging station At any moment Availability, Indicates the maximum capacity of the charging station; Based on power prediction and task progress assessment, robots with low power are prioritized for dispatch to available charging stations, and a preliminary dispatch plan is generated based on task priority. ,in It refers to the scheduling scheme; The initial task adjustment and charging scheduling decisions are sent to the robot unit to execute the scheduling instructions.

6. The inspection robot charging control system according to claim 1, characterized in that, The steps for outputting a global scheduling decision include: Each task is assigned a priority parameter. The priority parameter is divided into three values: high, medium, and low. ,middle ,Low ; A global scheduling decision model is constructed, defining the objective function as: minimizing task interruption time and charging wait time, while maximizing the continuity of high-priority tasks, i.e. The constraints also include robot power constraints, charging station capacity constraints, and task priority constraints. On edge computing nodes, preliminary scheduling decisions are made based on real-time data, and robots are assigned to charging stations. Optimization is then performed based on current battery level, task progress, and charging station availability. The initial decisions and data from edge computing nodes are uploaded to the cloud platform for comprehensive global analysis and optimization. Global optimization of the RL model is performed on a cloud platform using reinforcement learning (RL) algorithms. The state, action, reward, and objective function of the RL model are defined as follows: state The current system status, including robot battery level, task status, and charging station availability; action Scheduling decisions under the current state, including task allocation and charging plans; award Immediate reward calculated based on the objective function: ; Maximize cumulative reward : ,in Discount factor; The scheduling scheme optimized by RL on the cloud platform is distributed to each robot through edge computing nodes to execute scheduling instructions; Edge computing nodes and cloud platforms monitor execution in real time and dynamically adjust scheduling plans to deal with emergencies.

7. A charging device for a charging control system of an inspection robot, characterized in that, A charging control system for an inspection robot as described in any one of claims 1-6 is provided. The charging device is arranged based on the characteristics of the inspection task and the robot's movement trajectory, and dynamically adjusts the position and number of charging points based on the IoT data analysis results from the data analysis module. The charging device includes: Fixed charging stations include: The charging management unit monitors the charging process; The communication unit communicates with the robot unit and the cloud platform to receive and execute charging scheduling instructions. Mobile charging stations, including: The positioning and navigation unit actively moves to the vicinity of the robot to recharge based on the robot's location and charging needs; The charging coordination unit coordinates the charging time and sequence with multiple robots.

8. The charging device of the inspection robot charging control system according to claim 7, characterized in that, The charging device is arranged as follows: Collect historical inspection task data, analyze the frequency, location and path of inspection tasks, and output key areas and high-frequency inspection paths of inspection tasks. Based on the task analysis results, initial fixed charging points are set up in the inspection area, covering areas with frequent tasks and path intersections. IoT sensors are deployed in the inspection area to monitor the robot's location and battery level in real time, and upload the data to edge computing nodes and cloud platforms in real time. The location and number of charging points are dynamically adjusted as follows: Real-time monitoring of the usage of each fixed charging point, and statistics on charging frequency, charging duration and congestion; Based on real-time and historical data, the location and number of charging points are dynamically adjusted; temporary charging points are added in areas with high charging demand. The mobile charging station is equipped with a navigation system that actively heads to the robot's location to charge based on the robot's position and battery status.

Images