Intelligent agent control method and system

Abstract

The invention relates to the field of crane management, and discloses an intelligent agent control method and system.The method comprises the steps that operation parameters and operation environment data are collected in real time through multiple sensors arranged on a crane and subjected to standardization processing; transmitting the processed data to an operation state analysis model, and outputting an operation state evaluation value so as to analyze the current operation state of the crane and predict potential risks; based on the analysis result, an optimization control instruction is generated and issued to each execution mechanism controller, and corresponding mechanisms are driven to execute precise actions; through multi-source data fusion and intelligent analysis, real-time monitoring and risk assessment of the running state of the crane are achieved, the self-adaptive capacity and safety of the control system are improved, the manual intervention requirement is effectively reduced, and the method is suitable for intelligent crane control in the complex working environment.

Description

Technical Field

[0001] This invention relates to the field of crane management, and more specifically to a control method and system for an intelligent agent. Background Technology

[0002] In modern industry, construction, and logistics, cranes are core lifting equipment, and their operational safety and efficiency directly impact work processes. Traditional crane control relies mainly on manual experience or preset programs, which still has some shortcomings.

[0003] For example, data obtained from only a few sensors such as weight and position cannot capture multidimensional risks in complex environments in real time, such as the impact of sudden wind speed changes and dynamic movement of obstacles on the stability of cranes, resulting in a lag in risk assessment.

[0004] At the same time, without establishing a dynamic interaction model between operating parameters such as lifting weight and boom angle and environmental data such as wind speed and temperature, it is difficult to quantify the risk amplification effect caused by the coupling of multiple factors, such as the synergistic effect of strong winds exacerbating structural loads under heavy loads. Summary of the Invention

[0005] The purpose of this invention is to provide a control method and system for intelligent agents to solve the above-mentioned technical problems.

[0006] The objective of this invention can be achieved through the following technical solutions: A control method and system for an intelligent agent, comprising: S1. By using multiple sensors installed on the crane, the crane's operating parameters are collected in real time, and the working environment data is also collected. S2. After standardizing the collected operating parameters and working environment data, the data is transmitted to the operating status analysis model for analysis. Based on the operating status evaluation value output by the operating status analysis model, the current operating status of the crane is analyzed and evaluated, and operating risks are predicted. S3. Generate corresponding control commands based on the analysis and evaluation results; S4. The control commands are transmitted to the controllers of each actuator of the crane. Each actuator controller controls the corresponding actuator to perform the action according to the received commands.

[0007] In a further embodiment, the operating parameters include the lifting weight, boom angle, trolley position, crane position, and the operating speed of each actuator of the crane; The operational environment data includes wind speed, wind direction, temperature, and information on obstacles around the work site; The expression for the operational status analysis model is: ; In the formula, Indicates the first Item running parameters, Indicates the first Item of working environment parameters, , These represent the total number of parameters for operation parameters and operating environment data, respectively. , The weighting coefficients are determined based on historical data analysis. , These represent the measured values ​​of any one of the operating parameters and the working environment parameters, respectively. This indicates an indicator of the interaction between operating parameters and operational environment data. This represents the operational status assessment value.

[0008] A further solution involves the interaction and influence indicators between the operating parameters and the working environment data. The expression is: ; In the formula, - This indicates the start and end points of a monitoring period within a unit. Indicates the first The curves showing the change of each operating parameter over time. Indicates the first Curves showing the change of various working environment parameters over time. Indicates the first The first operating parameter and the second The curve showing the change of the interaction coefficient of the first working environment parameter over time. The first operating parameter and the second Interaction coefficients of individual work environment parameters It is obtained through a pre-trained recurrent neural network model. , The influencing factor is obtained through analysis of historical data.

[0009] A further approach, the method for analyzing and evaluating the current operating status of the crane, is as follows: The operational status evaluation value output by the operational status analysis model Compared with the preset lower threshold Upper limit threshold Compare and classify the operating status levels: Status level = .

[0010] A further proposed method for predicting operational risks is as follows: Interaction Influence Indicators of Operating Parameters and Working Environment Data Compared with the operating status assessment value Construct a risk prediction expression: ; In the formula, Indicates risk indicators, This represents the weighting factor, with a value range of 0 to 1; Historical risk indicators are input into a pre-trained long short-term memory network to obtain predicted risk values. ;Predict risk value With risk threshold range Comparison: when If so, the predicted operational risk is low; when If so, the predicted operational risk is medium; when If so, the predicted operational risk is high.

[0011] A further scheme, the weighting factor The dynamic adjustment rules are as follows: Under normal operating conditions The value is set to 0.5, taking into account the risk contributions of S and the interaction indicator R in a balanced way; When at least one work environment parameter is detected to exceed the safety threshold, the threshold is automatically increased. Up to 0.8, focusing on operational status assessment values. Risk weights.

[0012] A further proposed approach, based on the analysis and evaluation results, is to generate corresponding control commands using the following method: S31. Based on the combination of operating status level and predicted risk level, confirm the current operating condition level; when and When, the current operating condition level is excellent; when or When, the current operating condition level is good; when and If so, the current operating condition level is poor; S32. When the current operating condition level is excellent, maintain operation and conduct regular inspections according to the preset frequency; when the current operating condition level is good, adjust operating parameters and prevent risks; when the current operating condition level is poor, shut down the machine in an emergency and activate protection.

[0013] Further solutions, including methods for generating corresponding control commands based on analysis and evaluation results, also include: S33. When any of the following parameters exceeds the safety threshold, regardless of the current operating condition level, the targeted control command shall be executed first; including: If the lifting weight exceeds 100% of the rated value: immediately stop the lifting operation, while allowing a slow descent, with the descent speed limited to the lowest setting; If the wind speed exceeds the safety threshold: the crane boom will automatically retract to a safe angle of 45° and any rotation will be prohibited; If the obstacle is less than the safe distance: control the corresponding mechanism to move in the opposite direction until the obstacle is greater than the safe distance.

[0014] A control system for an intelligent agent, comprising: The multi-source data acquisition module uses various sensors installed on the crane to collect the crane's operating parameters in real time, while also collecting data on the working environment. The operation status analysis and risk prediction module is used to standardize the collected operation parameters and working environment data, and then transmit them to the operation status analysis model for analysis. Based on the operation status evaluation value output by the operation status analysis model, the module analyzes and evaluates the current operation status of the crane and predicts the operation risks. The instruction generation module is used to generate corresponding control instructions based on the analysis and evaluation results. The instruction execution module is used to transmit control instructions to the controllers of each actuator of the crane. Each actuator controller controls the corresponding actuator to perform actions according to the received instructions.

[0015] The beneficial effects of this invention are: (1) By integrating operating parameters and working environment data, a multi-dimensional analysis model is constructed to comprehensively evaluate the current status and potential risks of the crane; it can not only independently analyze the impact of a single parameter or environmental factor, but also quantify the synergistic effect between different factors, avoid the one-sidedness of traditional single-index evaluation, significantly improve the comprehensiveness and accuracy of the judgment of operating status, and provide a more reliable basis for safe operation. (2) By using recurrent neural networks to generate the interaction coefficients between parameters and the environment in real time, the impact of factors such as changes in operating conditions and environmental fluctuations on risk during equipment operation can be dynamically captured. Combined with the risk prediction model, risk trends can be perceived in advance and graded early warnings can be triggered, such as maintaining operation when the risk is low and forcing shutdown when the risk is high. This allows the control strategy to be automatically adjusted according to real-time changes, enhancing the adaptability to complex operating conditions, reducing the frequency of manual intervention, and improving response efficiency and safety. Attached Figure Description

[0016] The invention will now be further described with reference to the accompanying drawings.

[0017] Figure 1 This is a diagram illustrating the method steps of the present invention; Figure 2 This is a system structure diagram of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0019] Please see Figures 1-2 As shown, the present invention is a control method and system for an intelligent agent, comprising: S1. Utilize various sensors installed on the crane to collect the crane's operating parameters in real time, while also collecting operational environment data; the sensors include a weight sensor (lifting weight), an angle encoder (crane angle), a position sensor (trolley / crane position), a speed sensor (actuator speed), an anemometer (wind speed / direction), a temperature sensor, and a lidar / vision sensor (obstacle distance). S2. After standardizing the collected operating parameters and working environment data, the data is transmitted to the operating status analysis model for analysis. Based on the operating status evaluation value output by the operating status analysis model, the current operating status of the crane is analyzed and evaluated, and the operating risks are predicted. Standardization refers to filtering the raw data, such as Kalman filtering, or normalization, which is preprocessing by scaling to the 0-1 range, to eliminate the interference of dimensional differences on the model analysis. S3. Generate corresponding control commands based on the analysis and evaluation results; for example: command generation logic: safe state: generate "maintain current operation" command, and trigger periodic inspection (e.g., once every 30 minutes); warning state: generate "adjust operating parameters" command (e.g., reduce lifting speed by 30%), and issue a warning through the HMI interface or audible and visual alarm; dangerous state: generate "emergency stop" command, cut off power supply and lock brake.

[0020] S4. The control commands are transmitted to the controllers of each actuator of the crane. Each actuator controller controls the corresponding actuator to perform the action according to the received command. For example: Lifting / lowering action: The motor speed is adjusted by the frequency converter, and the analog signal (4-20mA or 0-10V) is received from the PLC. Crane boom retraction / slewing: driven by a servo motor (driving a reducer, receiving position / speed commands via PROFINET bus); Emergency shutdown: Hardware-level response is implemented through a safety relay module, ensuring that the main circuit power is cut off within 50ms after the command is issued. The above command execution steps are all existing technologies; implementation is sufficient.

[0021] By employing multiple types of sensors to achieve real-time multi-dimensional data acquisition, the system covers crane operating status and environmental risk factors, avoiding the limitations of traditional single-parameter monitoring. It integrates operating parameters such as lifting weight and boom angle with environmental data such as wind speed and obstacle distance, covering complex risks that cannot be captured by a single parameter, such as structural overload caused by "heavy load + strong wind," thus improving the comprehensiveness of risk assessment. Through a closed loop of "data acquisition-analysis-command-execution," the system automates the entire crane control process, reducing manual intervention and improving operational safety and efficiency in complex environments. In high-risk scenarios such as ports and construction sites, it can independently complete environmental perception and risk response, avoiding accidents caused by human misjudgment. By outputting assessment values ​​in real time through an operating status analysis model and combining them with a risk prediction model, it provides early warnings of potential risks, allowing for earlier intervention than manual monitoring.

[0022] The operating parameters include the lifting weight, boom angle, trolley position, crane position, and the operating speed of each actuator of the crane; The operational environment data includes wind speed, wind direction, temperature, and information on obstacles around the work site; The expression for the operational status analysis model is: ; In the formula, Indicates the first Item running parameters, Indicates the first Item of working environment parameters, , These represent the total number of parameters for operation parameters and operating environment data, respectively. , The weighting coefficient is determined based on historical data analysis, through regression analysis of historical data, such as least squares method or machine learning training, and reflects the importance of the parameter, such as lifting weight having a higher weight than temperature. , These represent the measured values ​​of any one of the operating parameters and the working environment parameters, respectively. This indicates an indicator of the interaction between operating parameters and operational environment data. This represents the operational status assessment value.

[0023] In this invention, the expression for the operational status analysis model is: Among them, the molecular part comprehensively considers operating parameters ( ) and environmental parameters ( The weighted sum of the two types of data reflects the independent impact of the two types of data on the operating status, while the denominator quantifies the synergistic effect between parameters, such as how strong winds exacerbate structural loads under heavy loads, avoiding the neglect of coupling risks due to the superposition of single factors; the above multi-dimensional fusion analysis achieves the comprehensiveness of the status assessment, for example, by simultaneously considering the joint impact of the crane boom angle and wind speed on stability; at the same time, by training the weight coefficients with historical data, the model can adapt to different working conditions and improve the adaptability of the assessment.

[0024] The interaction influence index between operating parameters and working environment data The expression is: ; In the formula, - This indicates the start and end points of a monitoring period within a unit. Indicates the first The curves showing the change of each operating parameter over time. Indicates the first Curves showing the change of various working environment parameters over time. Indicates the first The first operating parameter and the second The curve showing the change of the interaction coefficient of the first working environment parameter over time. The first operating parameter and the second Interaction coefficients of individual work environment parameters It is obtained through a pre-trained recurrent neural network model. , The influencing factor is obtained through analysis of historical data and determined through data statistics. It can be set to 1 or 2 to amplify / reduce the nonlinear effects of parameter changes, such as the effect of the square of wind speed on the overturning moment of a crane.

[0025] In this invention, unit time is captured by integration. - The cumulative effect of intrinsic parameter interactions reflects the dynamic changes of risk over time. This is achieved by using a pre-trained recurrent neural network model to obtain the first... The first operating parameter and the second Interaction coefficients of individual work environment parameters This approach leverages the characteristics of time-series data to learn the nonlinear dynamic relationships between parameters, such as the instantaneous impact of sudden wind speed changes on the stability of the crane boom angle. Using the aforementioned formulas, it can dynamically capture real-time coupling risks between parameters, for example, by calculating the superimposed effect of the "lifting weight - wind speed" interaction on the structural load in real time. Simultaneously, the recurrent neural network (RNN) model adapts to temporal changes in operating conditions, avoiding the shortcomings of traditional static models, such as fixed thresholds, which cannot reflect dynamic risks.

[0026] The method for analyzing and evaluating the current operating status of a crane is as follows: The operational status evaluation value output by the operational status analysis model Compared with the preset lower threshold Upper limit threshold Compare and classify the operating status levels: Status level = .

[0027] The method for predicting operational risks is as follows: Interaction Influence Indicators of Operating Parameters and Working Environment Data Compared with the operating status assessment value Construct a risk prediction expression: ; In the formula, Indicates risk indicators, This represents the weighting factor, with a value range of 0 to 1; Historical risk indicators are input into a pre-trained long short-term memory network to obtain predicted risk values. ;Predict risk value With risk threshold range Comparison: when If so, the predicted operational risk is low; when If so, the predicted operational risk is medium; when If so, the predicted operational risk is high.

[0028] The weighting factor The dynamic adjustment rules are as follows: Under normal operating conditions The value is set to 0.5, taking into account the risk contributions of S and the interaction indicator R in a balanced way; When at least one work environment parameter is detected to exceed the safety threshold, the threshold is automatically increased. Up to 0.8, focusing on operational status assessment values. Risk weights; The adjustment rule is implemented through a fuzzy logic controller, with real-time environmental parameters as input and a corresponding output. The dynamic correction value. When environmental parameters are normal. Maintain the baseline value; when environmental parameters approach the threshold, Increase slowly; when environmental parameters exceed the threshold... Rapidly increase; adjust in real time using fuzzy logic When environmental parameters fluctuate, the weights of the current state (S) and historical interaction risks (R) are automatically balanced to avoid lag or oversensitivity caused by fixed weights.

[0029] In this invention, the temporal characteristics of historical risk data are learned through an LSTM model to predict future risk trends in advance, such as risk changes in the next 10-30 minutes. Compared with traditional threshold alarms that only make real-time judgments, this allows for earlier intervention, playing a proactive and dynamic role in risk prediction. For example, before a typhoon arrives, the correlation between historical wind speed and state assessment values ​​can be used to predict that a crane is about to enter a high-risk state, triggering protective actions such as boom retraction in advance.

[0030] Multi-factor weighted dynamic balancing effect: weighting factors The system automatically adjusts based on the scenario; in high-risk environments such as those with excessive wind speeds, the current state assessment value is emphasized. This approach balances state and interaction risks (R) in typical scenarios, avoiding misjudgments caused by fixed weights. In contrast to traditional solutions, which often employ fixed weights or single-factor thresholds, this approach... Dynamic adjustments improve the accuracy of risk assessment in scenarios where the environment is stable but the equipment condition is deteriorating. The operational effectiveness of risk levels: dividing continuous risk values ​​into discrete levels (low / medium / high) and corresponding to different control strategies (such as maintaining operation, adjusting parameters, and emergency shutdown) improves human-machine collaboration efficiency and reduces invalid alarms.

[0031] The method for generating corresponding control commands based on the analysis and evaluation results is as follows: S31. Based on the combination of operating status level and predicted risk level, confirm the current operating condition level; when and When the current operating condition level is excellent; for example: generating a "normal operation" command and sending it to the PLC master station via industrial Ethernet; triggering a timed inspection task, such as performing sensor self-calibration every 30 minutes, and determining whether there is an abnormality by comparing redundant sensor data); when or When, the current operating condition level is good; when and If so, the current operating condition level is poor; S32. When the current operating condition level is excellent, maintain operation and conduct regular inspections according to the preset frequency; when the current operating condition level is good, adjust operating parameters and prevent risks; when the current operating condition level is poor, shut down the machine in an emergency and activate protection.

[0032] In this invention, a two-dimensional combined assessment of state and risk is used to classify operating conditions into three levels (excellent / good / poor), achieving a precise mapping between "risk level" and "control intensity," avoiding excessive intervention or insufficient response caused by traditional "one-size-fits-all" control strategies. Compared to traditional control systems that rely heavily on single threshold alarms, which are prone to problems such as "frequent alarms without substantial risk" or "risk accumulation without timely triggering," this invention maintains normal operation under excellent conditions, requiring only periodic inspections to reduce unnecessary intervention. Under poor conditions, it triggers emergency shutdown and multiple protection mechanisms, shortening the response time to critical risks and making it faster than manual intervention. Multi-dimensional risk collaborative handling considers both current conditions (e.g., overload) and potential risks (e.g., rising wind speed trends), avoiding misjudgments caused by single-dimensional assessments. For example, when the lifting weight is in a warning state and not exceeded, but wind speed is predicted to rise rapidly, the system preemptively treats it as a "poor" condition to prevent accidents caused by compound risks.

[0033] Methods for generating corresponding control commands based on analysis and evaluation results also include: S33. When any of the following parameters exceeds the safety threshold, regardless of the current operating condition level, the targeted control command shall be executed first; including: If the lifting weight exceeds 100% of the rated value: immediately stop the lifting operation, while allowing a slow descent, with the descent speed limited to the lowest setting; If the wind speed exceeds the safety threshold: the crane boom will automatically retract to a safe angle of 45° and any rotation will be prohibited; If the obstacle is less than the safe distance: control the corresponding mechanism to move in the opposite direction until the obstacle is greater than the safe distance.

[0034] In this invention, for extreme working conditions characterized by "danger and high risk," the risk of accidents is reduced to less than 1 / 5 of that of traditional single protection measures through the coordinated control of multiple actuators. When an overload of 110% is detected and the wind speed exceeds the safety threshold, the lifting action is automatically stopped and the boom is retracted to a safe angle to avoid overturning accidents caused by structural instability. Through differentiated control commands, such as limiting the slewing speed but allowing the boom to retract, equipment damage is minimized while ensuring safety. For example, limiting the slewing speed can reduce inertial forces and prevent the boom from colliding with surrounding obstacles; allowing the boom to retract can actively lower the center of gravity and improve overall stability. Audible and visual alarm signals are triggered synchronously with the automatic actions of the equipment, allowing the operator to obtain risk information within 10 seconds. Pre-set control logic reduces human error. For example, in the event of an emergency shutdown, the system automatically cuts off non-safety-related operation permissions to prevent the operator from accidentally pressing the restart button.

[0035] A control system for an intelligent agent, comprising: The multi-source data acquisition module uses various sensors installed on the crane to collect the crane's operating parameters in real time, while also collecting data on the working environment. The operation status analysis and risk prediction module is used to standardize the collected operation parameters and working environment data, and then transmit them to the operation status analysis model for analysis. Based on the operation status evaluation value output by the operation status analysis model, the module analyzes and evaluates the current operation status of the crane and predicts the operation risks. The instruction generation module is used to generate corresponding control instructions based on the analysis and evaluation results. The instruction execution module is used to transmit control instructions to the controllers of each actuator of the crane. Each actuator controller controls the corresponding actuator to perform actions according to the received instructions.

[0036] It should be noted that the calculation formulas and all parameters involved in the calculations in this invention have been dimensionless beforehand. The process of dimensionless processing is well known in the industry and will not be described here.

[0037] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A control method and system for an intelligent agent, characterized in that, include: S1. By using multiple sensors installed on the crane, the crane's operating parameters are collected in real time, and the working environment data is also collected. S2. After standardizing the collected operating parameters and working environment data, the data is transmitted to the operating status analysis model for analysis. Based on the operating status evaluation value output by the operating status analysis model, the current operating status of the crane is analyzed and evaluated, and operating risks are predicted. S3. Generate corresponding control commands based on the analysis and evaluation results; S4. The control commands are transmitted to the controllers of each actuator of the crane. Each actuator controller controls the corresponding actuator to perform the action according to the received commands.

2. The control method for an intelligent agent according to claim 1, characterized in that, The operating parameters include the lifting weight, boom angle, trolley position, crane position, and the operating speed of each actuator of the crane; The operational environment data includes wind speed, wind direction, temperature, and information on obstacles around the work site; The expression for the operational status analysis model is: ; In the formula, Indicates the first Item running parameters, Indicates the first Item of working environment parameters, , These represent the total number of parameters for operation parameters and operating environment data, respectively. , The weighting coefficients are determined based on historical data analysis. , These represent the measured values ​​of any one of the operating parameters and the working environment parameters, respectively. This indicates an indicator of the interaction between operating parameters and operational environment data. This represents the operational status assessment value.

3. The control method for an intelligent agent according to claim 2, characterized in that, The interaction influence index between operating parameters and working environment data The expression is: ； In the formula, - This indicates the start and end points of a monitoring period within a unit. Indicates the first The curves showing the change of each operating parameter over time. Indicates the first Curves showing the change of various working environment parameters over time. Indicates the first The first operating parameter and the second The curve showing the change of the interaction coefficient of the first working environment parameter over time. The first operating parameter and the second Interaction coefficients of individual work environment parameters It is obtained through a pre-trained recurrent neural network model. , The influencing factor is obtained through analysis of historical data.

4. The control method for an intelligent agent according to claim 2, characterized in that, The method for analyzing and evaluating the current operating status of a crane is as follows: The operational status evaluation value output by the operational status analysis model Compared with the preset lower threshold Upper limit threshold Compare and classify the operating status levels: Status level = .

5. The control method for an intelligent agent according to claim 4, characterized in that, The method for predicting operational risks is as follows: Interaction Influence Indicators of Operating Parameters and Working Environment Data Compared with the operating status assessment value Construct a risk prediction expression: ; In the formula, Indicates risk indicators, This represents the weighting factor, with a value range of 0 to 1; Historical risk indicators are input into a pre-trained long short-term memory network to obtain predicted risk values. ;Predict risk value With risk threshold range Comparison: when If so, the predicted operational risk is low; when If so, the predicted operational risk is medium; when If so, the predicted operational risk is high.

6. The control method for an intelligent agent according to claim 5, characterized in that, The weighting factor The dynamic adjustment rules are as follows: Under normal operating conditions The value is set to 0.5, taking into account the risk contributions of S and the interaction indicator R in a balanced way; When at least one work environment parameter is detected to exceed the safety threshold, the threshold is automatically increased. Up to 0.8, focusing on operational status assessment values. Risk weights.

7. The control method for an intelligent agent according to claim 5, characterized in that, The method for generating corresponding control commands based on the analysis and evaluation results is as follows: S31. Based on the combination of operating status level and predicted risk level, confirm the current operating condition level; when and When, the current operating condition level is excellent; when or When, the current operating condition level is good; when and If so, the current operating condition level is poor; S32. When the current operating condition level is excellent, maintain operation and conduct regular inspections according to the preset frequency; when the current operating condition level is good, adjust operating parameters and prevent risks; when the current operating condition level is poor, shut down the machine in an emergency and activate protection.

8. The control method for an intelligent agent according to claim 7, characterized in that, Methods for generating corresponding control commands based on analysis and evaluation results also include: S33. When any of the following parameters exceeds the safety threshold, regardless of the current operating condition level, the targeted control command shall be executed first; including: If the lifting weight exceeds 100% of the rated value: immediately stop the lifting operation, while allowing a slow descent, with the descent speed limited to the lowest setting; If the wind speed exceeds the safety threshold: the crane boom will automatically retract to a safe angle of 45° and any rotation will be prohibited; If the obstacle is less than the safe distance: control the corresponding mechanism to move in the opposite direction until the obstacle is greater than the safe distance.

9. A control system for an intelligent agent, characterized in that, The control system is used to implement the control method for the intelligent agent as described in claim 1, including: The multi-source data acquisition module uses various sensors installed on the crane to collect the crane's operating parameters in real time, while also collecting data on the working environment. The operation status analysis and risk prediction module is used to standardize the collected operation parameters and working environment data, and then transmit them to the operation status analysis model for analysis. Based on the operation status evaluation value output by the operation status analysis model, the module analyzes and evaluates the current operation status of the crane and predicts the operation risks. The instruction generation module is used to generate corresponding control instructions based on the analysis and evaluation results. The instruction execution module is used to transmit control instructions to the controllers of each actuator of the crane. Each actuator controller controls the corresponding actuator to perform actions according to the received instructions.

Images