Intelligent robot operation management system based on electric power field

Abstract

The invention discloses an intelligent robot operation management system based on an electric power field, relates to the technical field of intelligent robot operation management, and solves the problems that a robot in the electric power field cannot accurately recognize obstacles temporarily appearing in the field, and the collision risk is easily increased. According to the invention, the obtained environment information is preprocessed, and the deep learning algorithm is combined to carry out target detection and identification, so that the system can accurately distinguish static and dynamic target objects in the environment; on the basis, the system further applies a time-to-collision model and a collision distance calculation model to carry out detailed collision risk analysis on the dynamic target object, the system obtains a risk score of the target object by combining factors such as different target object types, relative speeds and distances, and carries out grading marking on the target object according to the score; according to the grading marks, the robot can adopt corresponding strategies when facing target objects with different risk grades, so that possible collision accidents are effectively avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent robot operation management, and in particular to an intelligent robot operation management system based on a power site. Background Art

[0002] Power field operations are gradually shifting from traditional manual operations to more intelligent robot systems. Power field refers to the workplaces where power facilities such as substations and transmission lines are located. These places have complex working environments and high risks, so higher requirements are placed on the intelligent management of operations. The application of intelligent robots in power field includes inspection, operation, maintenance and other tasks. Through advanced sensor technology, machine vision, path planning algorithms and automated control technology, intelligent robots can autonomously complete various operations in dangerous environments such as high voltage and strong magnetic fields, effectively improving the safety and operation efficiency of the power system.

[0003] The existing intelligent robot operation management system in power fields has improved the automation level of operations to a certain extent, but in actual applications it is still not accurate enough in identifying environmental changes, especially in identifying and handling sudden obstacles.

[0004] The current power site intelligent robot operation management system still has obvious deficiencies in dealing with dynamic changes in the on-site environment, especially in the identification and avoidance of some dynamic targets and sudden obstacles. In order to solve these problems, it is necessary to further research and develop an intelligent management system that can perceive environmental changes in real time and make dynamic adjustments to ensure the safety and efficiency of power site operations. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an intelligent robot operation management system based on a power site, which solves the problems in the background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent robot operation management system based on a power site, comprising:

[0007] The information acquisition module is used to initially acquire environmental information around the robot and send the acquired environmental information data to the subsequent information processing module, wherein the environmental information includes terrain, dynamic and static target information;

[0008] The information processing module is used to pre-process the received real-time collected environmental information, including filtering and denoising, to reduce sensor errors and environmental noise. After that, the three-dimensional point cloud model of the surrounding environment is generated through the lidar data, and the two-dimensional and three-dimensional features of the target object are generated using the image obtained by the camera combined with the depth information. Then, based on the pre-processed data, the camera image is used to detect the target, identify and classify the target objects in the environment using a deep learning algorithm. For more accurate target recognition, semantic segmentation technology can be used to extract the boundaries and contours of each object.

[0009] A risk assessment module is used to perform risk assessment on the target object according to the data results preprocessed by the information processing module, and mark the risk level of the corresponding target object according to the result of the risk assessment;

[0010] The analysis module is used to set a safety range for the robot. Based on the evaluation results of the risk assessment module, the module determines whether the target object entering the range is a non-temporary target object or a temporary target object, and takes corresponding measures based on the judgment results.

[0011] The machine execution module is used to enable the robot to make corresponding action strategies based on the judgment results of the analysis module.

[0012] Preferably, in the information collection module, the specific method of obtaining is:

[0013] By using multi-sensor fusion technology, appropriate sensors are selected and arranged on the robot to obtain all-round environmental information; the sensors include: lidar, camera, radar, ultrasonic sensor, and IMU composed of accelerometer and gyroscope.

[0014] Preferably, in the risk assessment module, the specific method of risk assessment is:

[0015] S1: Determine all targets on the driving path;

[0016] S2: Distinguish static and dynamic objects by comparing between frames;

[0017] S3: Get the dynamic target, analyze the dynamic target again, determine whether a collision occurs, and use the time-to-collision model mathematical model to specifically calculate the collision probability between the robot and the target: Then, perform risk assessment based on the collision probability using the following formula:

[0018] S4: Classify and label the dynamic corresponding targets according to the risk assessment value.

[0019] Preferably, in step S3, it specifically includes:

[0020] Time-to-collision TTC is calculated by the following formula:

[0021]

[0022] Among them, v rel is the relative velocity vector between the target and the robot: v rel =v o -v r , v r is the robot velocity vector, v o is the velocity vector of the target object, and d is the relative position vector between the target object and the robot: d = p o -p r , p r is the robot’s current position vector, p o is the current position vector of the target object;

[0023] If TTC>0, it means there is a possibility of collision, and if TTC≤0, it is considered that no collision will occur;

[0024] Then the collision distance is calculated. After the time TTC, whether the distance between the two is less than the safe distance, the collision distance can be calculated by the following formula:

[0025] d collision =|d+TTC·v rel |

[0026] Among them, d collision is the collision distance, if: d collision ≤(R r +R o ), it means that the robot and the target will collide after the time TTC, where: Rr is the robot radius or the minimum safe distance, Ro: is the target radius or the minimum safe distance, and the specific parameter value is set by the staff;

[0027] Combined with the above, the collision probability between the robot and the target object is determined by the following formula:

[0028]

[0029] Among them, P collision represents the probability of detected collision;

[0030] Next, risk assessment is performed based on the collision probability using the following formula:

[0031]

[0032] Among them, R score is the risk assessment value of the target object, w typeis the weight of the target type; α, β, γ are the weight coefficients of collision probability, distance and speed respectively; ∈ is a small constant to avoid the situation where the denominator is zero when the distance is zero; v max is the maximum value of the relative velocity, used for normalization.

[0033] Preferably, in step S4, it specifically includes:

[0034] When R score When G1 is less than the preset value and greater than 0, it is marked as a low-risk target. score When it is less than the preset value G2 and greater than G1, it is marked as a medium-risk target. score When the preset value G2 is exceeded, it is marked as a high-risk target. In other cases, it is marked as a static target, and G1 <G2。

[0035] Preferably, in the analysis module:

[0036] The method for judging the target object entering the safety range is as follows:

[0037] A safety range is set for the robot. During driving, the robot judges the surrounding information. When it is determined that an object has entered the range, the robot judges the target object. When a marked target object is identified to appear in the range, it is judged to be a non-temporary target object. When a marked target object is not identified and appears in the range, it is judged to be a sudden temporary target object.

[0038] Preferably, the judgment result is processed in the following manner:

[0039] When it is determined to be a non-temporary target:

[0040] If the target object is marked as low risk, the robot can continue to maintain the current driving path and speed, but will continue to monitor and make advance plans;

[0041] If the target is marked as medium risk and has a high probability of collision, the robot should slow down and prepare to take evasive measures. It will also update the position and speed of the target in real time and adjust the path to avoid collision.

[0042] If the target object is marked as high risk, the robot should immediately take avoidance actions, including emergency stops, path replanning, and even reversing measures.

[0043] Preferably, when it is determined to be a sudden temporary target:

[0044] If the sensor detects that a sudden temporary target object enters the safe range, the robot will immediately mark it as a special risk target and perform emergency processing.

[0045] The preferred emergency treatment method is:

[0046] The time to collision between a high-risk target and the robot is calculated using the following formula:

[0047]

[0048] Among them, d currentd is the current distance, v tel is the relative speed between the special risk object and the robot;

[0049] Calculate the surrounding environmental impact coefficient E based on the surrounding environmental factors env , the coefficient is determined by the following formula:

[0050]

[0051] Among them, N obstacles is the number of objects around the robot, S free is the area of ​​the drivable space, S tota is the total area within the robot’s current safety range;

[0052] Combine TTC1 with the environmental impact factor to calculate the comprehensive risk score R special :

[0053]

[0054] Then, according to R special The robot takes different actions depending on the value of

[0055] Preferably, the robot takes different actions:

[0056] If R special > The preset value H2 means that the special risk target is very close to the robot and the environment is complex. The robot should stop immediately to avoid any possible collision risk;

[0057] If the preset value is H1 <R special <Preset value H2, the robot tries to avoid and re-plan the path to avoid the special risk target. At this time, the robot may slow down or change the driving direction;

[0058] If R special <Default value H1, the environment is relatively simple and the special risk target is far away. The robot can continue to monitor its dynamics and move forward slowly while remaining vigilant.

[0059] The present invention provides an intelligent robot operation management system based on the power site. Compared with the prior art, it has the following beneficial effects:

[0060] (1) The present invention pre-processes the acquired environmental information and combines it with deep learning algorithms for target detection and recognition. The system can accurately distinguish between static and dynamic targets in the environment. On this basis, the system further applies the time-to-collision model (TTC) and the collision distance calculation model to perform detailed collision risk analysis on dynamic targets. By combining different target types, relative speeds, distances and other factors, the system obtains the risk score of the target and grades and labels the target according to the score. Such graded labeling enables the robot to adopt corresponding strategies when facing targets of different risk levels, thereby effectively avoiding possible collision accidents. This risk assessment mechanism based on real-time data ensures the safety of the robot when performing tasks in complex power sites, enabling it to actively avoid potential dangers and avoid damage to power equipment and operators. The system can also continuously adjust and optimize the robot's action strategy based on the risk assessment results, further improving the ability to respond to emergencies.

[0061] (2) The present invention, through the emergency handling mechanism of special risk targets, the system can quickly calculate the time to collision (TTC) between the special risk target and the robot, and combine it with the environmental impact coefficient to obtain a comprehensive risk score. According to the score result, the robot can immediately take emergency measures including emergency stop, re-planning of the path, etc., so as to avoid potential major accidents. This feature of the system is particularly suitable for working scenes such as power sites with complex environments and many emergencies, and can significantly reduce power outages or equipment damage accidents caused by emergencies. Through timely recording and subsequent learning of emergencies, the system can also continuously optimize its emergency handling strategy, making the robot more efficient and safer in future operations; BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The present invention will be further described below in conjunction with the accompanying drawings.

[0063] Figure 1 This is a structural framework diagram of an intelligent robot operation management system based on a power site according to the present invention;

[0064] Figure 2 The present invention is a risk assessment and determination schematic diagram of an intelligent robot operation management system based on a power site. DETAILED DESCRIPTION

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] Embodiment 1

[0067] See also Figure 1-Figure 2 ,The present invention provides an intelligent robot operation management system based on the power site, including an information collection module, an information processing module, a risk assessment module, an analysis module, and a robot execution module;

[0068] The information acquisition module is used to initially acquire environmental information around the robot and send the acquired environmental information data to the subsequent information processing module, wherein the environmental information includes terrain, dynamic and static target information;

[0069] The specific method of obtaining is: using multi-sensor fusion technology, selecting appropriate sensors and arranging them on the robot to obtain environmental information in all directions; the sensors include: laser radar (LiDAR), camera, radar, ultrasonic sensor, IMU;

[0070] Specifically, LiDAR generates three-dimensional point cloud data of the environment by emitting laser pulses and measuring their return time; the point cloud data can describe in detail the height changes and slopes of the surrounding terrain, and detect static or dynamic targets. Among them, LiDAR is suitable for accurately measuring the distance and size of surrounding objects, and can still maintain high accuracy at a long distance; when combined with time series data, LiDAR can also monitor the movement trajectory of dynamic objects through continuous scanning;

[0071] The camera is mainly used to collect two-dimensional images and video information of the environment. Through these images, the robot can identify and classify different objects, such as vehicles, people, signs, etc. Based on the image data of the camera, the robot can extract features such as color, shape, texture, etc. for object detection and classification;

[0072] Radar detects objects in the environment by emitting electromagnetic waves and receiving reflected signals; it can accurately measure the relative speed and distance of objects, especially in bad weather (such as rain and fog). The radar's penetrating ability enables it to provide key speed and direction information when detecting dynamic objects; by analyzing the reflected wave signal, radar can identify and track multiple targets, especially in long-distance detection, the radar's effectiveness and reliability are very high;

[0073] Ultrasonic sensors detect objects and targets at close range by emitting ultrasonic waves and receiving reflected signals. They are usually used to detect slow-moving objects and static targets at close range, such as walls and pillars. Ultrasonic sensors have weak reflections on soft or non-metallic materials, but can complement the detection blind spots of other sensors at short distances. Due to the wavelength characteristics of ultrasonic waves, they can identify surface features, such as the roughness of walls or small objects on the ground.

[0074] IMU consists of an accelerometer and a gyroscope, and is used to measure the robot's acceleration, angular velocity, and azimuth in real time. It can help the robot understand its own motion state (such as tilt, rotation, acceleration) and its response to external shock or vibration. Combined with other sensor data, IMU can provide motion compensation information, allowing the robot to maintain stable navigation and attitude control in complex terrain or high-speed movement.

[0075] By fusing the data from the above sensors, the robot can obtain a more comprehensive and reliable environmental perception; for example:

[0076] Combining LiDAR with camera: The combination of LiDAR's 3D point cloud data and the camera's 2D image can achieve high-precision object detection and classification, especially under complex lighting conditions. Combining the data of the two can make up for the shortcomings of a single sensor.

[0077] Combination of radar and IMU: The dynamic object velocity information provided by radar is combined with the posture data of IMU to more accurately predict the movement trajectory of the target object, especially when the robot itself is moving;

[0078] The information processing module is used to pre-process the received real-time collected environmental information, including filtering and denoising, to reduce sensor errors and environmental noise. After that, the 3D point cloud model of the surrounding environment is generated through the LiDAR data, and the 2D and 3D features of the target object are generated using the image obtained by the camera combined with the depth information;

[0079] Then, based on the preprocessed data, deep learning algorithms (such as YOLO and Mask R-CNN) are used to detect objects in the camera images, identify and classify objects in the environment (such as vehicles, people, and equipment). For more accurate object recognition, semantic segmentation techniques (such as U-Net) can be used to extract the boundaries and contours of each object.

[0080] The risk assessment module is used to conduct risk assessment on the target object according to the preprocessed data results, and mark the risk level of the corresponding target object according to the risk assessment results;

[0081] See also Figure 2 The specific methods of risk assessment are:

[0082] S1: Determine all targets on the driving path;

[0083] S2: Distinguish static and dynamic objects by comparing between frames;

[0084] Among them, inter-frame comparison is to judge whether the object has moved by comparing the position changes of the same object in adjacent frames (time series data). If the object keeps the same position between consecutive frames, the object is regarded as a static target; if the position of the object changes significantly, it is regarded as a dynamic target.

[0085] S3: Get the dynamic target, analyze the dynamic target again, determine whether a collision occurs, and use the time-to-collision (TTC) mathematical model to specifically calculate the collision probability between the robot and the target:

[0086] Time-to-collision TTC is calculated by the following formula:

[0087]

[0088] Among them, v rel is the relative velocity vector between the target and the robot: v rel =v o -v r , v r is the robot velocity vector, v o is the velocity vector of the target object, and d is the relative position vector between the target object and the robot: d = p o -p r , p r is the robot’s current position vector, p o is the current position vector of the target object;

[0089] If TTC>0, it means there is a possibility of collision, and if TTC≤0, it is considered that no collision will occur;

[0090] Then the collision distance is calculated. After the time TTC, whether the distance between the two is less than the safe distance, the collision distance can be calculated by the following formula:

[0091] d collision =|d+TTC·v rel |

[0092] Among them, d collision is the collision distance, if: d collision ≤(R r +R o ), it means that the robot and the target will collide after the time TTC, where: Rr is the robot radius or the minimum safe distance, Ro: is the target radius or the minimum safe distance, and the specific parameter value is set by the staff;

[0093] Combined with the above, the collision probability between the robot and the target object is determined by the following formula:

[0094]

[0095] Among them, P collision represents the probability of detected collision;

[0096] Next, risk assessment is performed based on the collision probability using the following formula:

[0097]

[0098] Among them, R score is the risk assessment value of the target object, w type is the weight of the target type, which can adjust the sensitivity of the score to different types of targets. For example, if a person is detected, a higher weight should be given because of the greater potential harm. α, β, and γ are the weight coefficients of collision probability, distance, and speed, respectively. The influence of these factors on the final score needs to be adjusted according to the application scenario. For example, in a scenario where speed is more important, γ should be set to a larger value. ∈ is a small constant to avoid the situation where the denominator is zero when the distance is zero. v max is the maximum value of the relative velocity, used for normalization;

[0099] S4: According to the risk assessment value R score The dynamic corresponding targets are graded and marked. score When G1 is less than the preset value and greater than 0, it is marked as a low-risk target. score When it is less than the preset value G2 and greater than G1, it is marked as a medium-risk target. score When the preset value G2 is exceeded, it is marked as a high-risk target. In other cases, it is marked as a static target, and G1 <G2;

[0100] The analysis module is used to set a safety range for the robot, determine whether the target object entering the range is a non-temporary target object or a temporary target object, and make corresponding processing according to the judgment result;

[0101] The method for judging the target object entering the safety range is as follows:

[0102] A safety range is set for the robot. During the driving process, the robot judges the surrounding information. When it is determined that an object has entered the range, the robot judges the target object. When the marked target object is identified to appear in the range, it is judged as a non-temporary target object. When the marked target object is not identified and appears in the range, it is judged as a sudden temporary target object.

[0103] The corresponding processing method for the judgment result is:

[0104] When it is determined to be a non-temporary target:

[0105] If the target object is marked as low risk, the robot can continue to maintain the current driving path and speed, but will continue to monitor and make advance plans;

[0106] If the target is marked as medium risk and has a high probability of collision, the robot should slow down and prepare to take evasive measures. It will also update the position and speed of the target in real time and adjust the path to avoid collision.

[0107] If the target object is marked as high risk, the robot should immediately take evasive action, including emergency stop, path replanning, or even reversing to ensure safety;

[0108] In the above method, the robot obtains environmental data through sensor fusion, uses deep learning to identify the target object, combines motion prediction and trajectory planning to evaluate the collision risk, and executes corresponding avoidance measures through decision trees or rule engines. At the same time, it uses real-time control and feedback loops to continuously optimize the action strategy. This is an existing technology and will not be elaborated here.

[0109] When it is determined to be a sudden temporary target:

[0110] If the sensor detects that a sudden temporary target object has entered the safe range, the robot will immediately mark it as a special risk target and perform emergency treatment;

[0111] Embodiment 2

[0112] In the specific implementation process, this embodiment is based on the first embodiment, and the difference from the first embodiment is that this embodiment specifically introduces the emergency treatment method of the special risk target object;

[0113] The emergency treatment methods are:

[0114] The time to collision between a high-risk target and the robot is calculated using the following formula:

[0115]

[0116] Among them, d currentd is the current distance, v tel is the relative speed between the special risk object and the robot;

[0117] The surrounding environment impact coefficient E is calculated based on factors such as the density of other targets in the surrounding environment, drivable space, and path complexity. env , the coefficient can be determined by the following formula:

[0118]

[0119] Among them, N obstacles is the number of objects around the robot, S freeis the area of ​​the drivable space, S tota is the total area within the robot’s current safety range;

[0120] Combine TTC1 with the environmental impact factor to calculate the comprehensive risk score R special :

[0121]

[0122] Then, according to R special The robot takes different actions:

[0123] If R special > The preset value H2 means that the special risk target is very close to the robot and the environment is complex. The robot should stop immediately to avoid any possible collision risk;

[0124] If the preset value is H1 <R special <Preset value H2, the robot tries to avoid and re-plan the path to avoid the special risk target. At this time, the robot may slow down or change the driving direction;

[0125] If R special <Default value H1, the environment is relatively simple and the special risk target is far away. The robot can continue to monitor its dynamics and move forward slowly while remaining vigilant;

[0126] After taking initial action, the robot will continue to monitor the changes in the special risk target and the surrounding environment. special If a significant change occurs, the robot should adjust its action strategy in a timely manner; the robot records the event for subsequent training of the machine learning model to continuously optimize the strategy for dealing with special risk targets;

[0127] The machine execution module is used to make the robot make corresponding action strategies according to the judgment results of the analysis module;

[0128] Embodiment 3

[0129] The specific implementation process of this embodiment includes the entire implementation process of the above two groups of embodiments.

[0130] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0131] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An intelligent robot operation management system based on power site, characterized in that: include: The information acquisition module is used to initially acquire environmental information around the robot and send the acquired environmental information data to the subsequent information processing module, wherein the environmental information includes terrain, dynamic and static target information; The information processing module is used to pre-process the received real-time collected environmental information, including filtering and denoising, to reduce sensor errors and environmental noise. After that, the three-dimensional point cloud model of the surrounding environment is generated through the lidar data, and the two-dimensional and three-dimensional features of the target object are generated using the image obtained by the camera combined with the depth information. Then, based on the pre-processed data, the camera image is used to detect the target, identify and classify the target objects in the environment using a deep learning algorithm. For more accurate target recognition, semantic segmentation technology can be used to extract the boundaries and contours of each object. A risk assessment module is used to perform risk assessment on the target object according to the data results preprocessed by the information processing module, and mark the risk level of the corresponding target object according to the result of the risk assessment; The analysis module is used to set a safety range for the robot. Based on the evaluation results of the risk assessment module, the module determines whether the target object entering the range is a non-temporary target object or a temporary target object, and takes corresponding measures based on the judgment results. The machine execution module is used to enable the robot to make corresponding action strategies based on the judgment results of the analysis module.

2. According to claim 1, an intelligent robot operation management system based on a power site is characterized in that: In the information collection module, the specific method of obtaining is: By using multi-sensor fusion technology, appropriate sensors are selected and arranged on the robot to obtain all-round environmental information; the sensors include: lidar, camera, radar, ultrasonic sensor, and IMU composed of accelerometer and gyroscope.

3. The intelligent robot operation management system based on the power site according to claim 2 is characterized in that: In the risk assessment module, the specific methods of risk assessment are: S1: Determine all targets on the driving path; S2: Distinguish static and dynamic objects by comparing between frames; S3: Get the dynamic target, analyze the dynamic target again, determine whether a collision occurs, and use the time-to-collision model mathematical model to specifically calculate the collision probability between the robot and the target: Then, perform risk assessment based on the collision probability using the following formula: S4: Classify and label the dynamic corresponding targets according to the risk assessment value.

4. The intelligent robot operation management system based on the power site according to claim 3 is characterized in that: In step S3, it specifically includes: Time-to-collision TTC is calculated by the following formula: Among them, v rel is the relative velocity vector between the target and the robot: v rel =v o -v r , v r is the robot velocity vector, v o is the velocity vector of the target object, and d is the relative position vector between the target object and the robot: d = p o -p r , p r is the robot’s current position vector, p o is the current position vector of the target object; If TTC>0, it means there is a possibility of collision, if TTC≤0, it is considered that no collision will occur; Then the collision distance is calculated. After the time TTC, whether the distance between the two is less than the safe distance, the collision distance can be calculated by the following formula: d collision =|d+TTC·v rel | Among them, d collision is the collision distance, if: d collision ≤(R r +R o ), it means that the robot and the target will collide after the time TTC, where: Rr is the robot radius or the minimum safe distance, Ro: is the target radius or the minimum safe distance, and the specific parameter value is set by the staff; Combined with the above, the collision probability between the robot and the target object is determined by the following formula: Among them, P collision represents the probability of detected collision; Next, risk assessment is performed based on the collision probability using the following formula: Among them, R score is the risk assessment value of the target object, w type is the weight of the target type; α, β, γ are the weight coefficients of collision probability, distance and speed respectively; ∈ is a small constant to avoid the situation where the denominator is zero when the distance is zero; v max is the maximum value of the relative velocity, used for normalization.

5. The intelligent robot operation management system based on the power site according to claim 4 is characterized in that: In step S4, it specifically includes: When R score When G1 is less than the preset value and greater than 0, it is marked as a low-risk target. score When it is less than the preset value G2 and greater than G1, it is marked as a medium-risk target. score When the preset value G2 is exceeded, it is marked as a high-risk target. In other cases, it is marked as a static target, and G1 <G2。 6. The intelligent robot operation management system based on the power site according to claim 1 is characterized in that: In the analysis module: The method for judging the target object entering the safety range is as follows: A safety range is set for the robot. During driving, the robot judges the surrounding information. When it is determined that an object has entered the range, the robot judges the target object. When a marked target object is identified to appear in the range, it is judged to be a non-temporary target object. When a marked target object is not identified and appears in the range, it is judged to be a sudden temporary target object.

7. The intelligent robot operation management system based on the power site according to claim 6 is characterized in that: The corresponding processing method for the judgment result is: When it is determined to be a non-temporary target: If the target object is marked as low risk, the robot can continue to maintain the current driving path and speed, but will continue to monitor and make advance plans; If the target is marked as medium risk and has a high probability of collision, the robot should slow down and prepare to take evasive measures. It will also update the position and speed of the target in real time and adjust the path to avoid collision. If the target object is marked as high risk, the robot should immediately take avoidance actions, including emergency stops, path replanning, and even reversing measures.

8. The intelligent robot operation management system based on the power site according to claim 5 is characterized in that: When it is determined to be a sudden temporary target: If the sensor detects that a sudden temporary target object enters the safe range, the robot will immediately mark it as a special risk target and perform emergency processing.

9. The intelligent robot operation management system based on the power site according to claim 8 is characterized in that: The emergency treatment methods are: Calculate the time until collision between a high-risk target and the robot using the following formula: Among them, d currentd is the current distance, v tel is the relative speed between the special risk object and the robot; Calculate the surrounding environmental impact coefficient E based on the surrounding environmental factors env , the coefficient is determined by the following formula: Among them, N obstacles is the number of objects around the robot, S free is the area of ​​the drivable space, S tota is the total area within the robot’s current safety range; Combine TTC1 with the environmental impact factor to calculate the comprehensive risk score R special : Then, according to R special The robot takes different actions depending on the value of 10. The intelligent robot operation management system based on the power site according to claim 9 is characterized in that: The robot takes different actions: If R special > The preset value H2 means that the special risk target is very close to the robot and the environment is complex. The robot should stop immediately to avoid any possible collision risk; If the preset value is H1 <R special <Preset value H2, the robot tries to avoid and re-plan the path to avoid the special risk target. At this time, the robot may slow down or change the driving direction; If R special <Default value H1, the environment is relatively simple and the special risk target is far away. The robot can continue to monitor its dynamics and move forward slowly while remaining vigilant.

Images