Anti-falling method and system of cable crane type coating robot

Abstract

This invention belongs to the field of mobile robot technology and discloses a method for preventing cable roll-type coating robots from detaching. The specific steps of this method are as follows: S1 Collect state parameters; S2 Calculate the score value; S3 Determine the risk and output control commands. This invention acquires multi-source disturbance data in real time by deploying tension sensors, attitude angle sensors, and triaxial accelerometers on key parts of the robot body, and constructs a structured state parameter vector, thereby achieving a comprehensive quantitative characterization of the operating state. Compared with traditional methods that rely on a single threshold to determine risk, this multi-dimensional information fusion mechanism can simultaneously perceive force, attitude, and dynamic disturbances, effectively improving the ability to identify and judge potential detachment risks. It is particularly suitable for dynamic stability monitoring in high-altitude flexible track scenarios, ensuring safe and continuous operation of the robot during long-distance operation.

Description

Technical Field

[0001] This invention belongs to the field of mobile robot technology, specifically a method and system for preventing the coating of cables on a cable-lined robot from falling off. Background Technology

[0002] Cable-mounted coating robots are widely used in high-altitude flexible track operations. They move on suspended cables using their own walking system and complete tasks such as surface treatment and spraying. However, due to the flexibility, swaying and instability of the cable structure, the robot is easily affected by wind load disturbances, tension fluctuations, posture imbalances and center of gravity shifts during operation, which can lead to high-risk events such as derailment and falls. Currently, the industry often enhances stability by increasing clamping force or designing limiting structures, but these technical means are mostly mechanical improvements and lack the ability to accurately perceive and dynamically respond to disturbances, making it difficult to adapt to the stable operation requirements in changing environments.

[0003] Meanwhile, traditional anti-detachment technologies generally rely on a single physical quantity threshold for judgment, such as triggering motion control when the tension is too high or the tilt angle exceeds the warning value. However, such methods ignore the synergistic relationship between multiple states, and the fixed threshold mechanism is prone to misjudgment when dealing with dynamic working conditions. Faced with increasingly complex operating environments, there is an urgent need to establish a stability recognition mechanism with multi-parameter fusion capabilities, high judgment sensitivity, and flexible response logic, and to form a closed-loop control path with real-time adjustment and feedback learning capabilities to effectively improve the anti-detachment capability and track adhesion stability of coating robots under multiple disturbances. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for preventing the coating of cables on a cable-carrying vehicle from falling off, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for preventing the cable coating robot from detaching, the specific steps of which are as follows:

[0006] S1 acquires state parameters: By using tension sensors, attitude angle sensors and three-axis accelerometers set on the robot body, the changes in cable tension, tilt angle, lateral acceleration and centroid offset distance are acquired in real time to construct a set of state parameters;

[0007] S2 calculates the score: The state parameter vector is input into the disturbance factor aggregation score function to obtain the score S reflecting the current operational stability of the robot;

[0008] S3 determines the risk and outputs control commands: It compares the score value S with the judgment threshold T dynamically calculated based on the sliding window. If the score value S is greater than the judgment threshold T, it initiates the corresponding control command and drives the execution device to complete the process in sequence; it increases the clamping force of the clamping device; and it adjusts the contact angle of the support structure.

[0009] Adjust the release angle of the tension relief device to suppress the risk of detachment and maintain trajectory adhesion stability.

[0010] Preferably, the specific steps for collecting state parameters in step S1 are as follows:

[0011] S11 deploys tension sensors, attitude angle sensors, and triaxial accelerometers at key operating points of the cable-mounted coating robot. The tension sensors are installed between the main load-bearing structures on the robot's load side to continuously monitor changes in cable stress during operation and output cable tension values ​​(in N). The attitude angle sensors are located at the geometric center of the robot body to acquire the tilt angle (in °) of the coating module relative to the horizontal reference plane in real time. The triaxial accelerometers are installed at the connection between the chassis and the active drive component to collect X / Y / Z acceleration changes (in m / s²) to assess the degree of trajectory disturbance and the trend of operational inertia. These sensors operate synchronously at a 100Hz sampling frequency, outputting standardized physical quantity streams as basic data channels for connection to the state modeling module.

[0012] Based on data collected by various sensors, the S12 system synchronously collects data including: real-time changes in cable tension (in Newtons), robot tilt angle (in degrees), and lateral acceleration (in m / s²). It further combines this with the relative offset distance of the center of gravity (in meters) measured by the center of gravity position detection device. These values ​​are quantified, time-aligned, and encoded to form a state parameter vector with a fixed structure. This vector serves as the input variable for the scoring function, reflecting characteristics such as the robot's operating posture, track adhesion, and disturbance amplitude, providing a quantitative basis for subsequent risk identification and control strategy generation.

[0013] The formula for calculating the relative offset distance of the centroid is:

[0014] ;

[0015] In the formula, Spatial offset distance from the robot's center of mass to the cable axis (unit: m);

[0016] The coordinates of the robot's center of mass in three-dimensional space;

[0017] : Coordinates of the cable axis reference point;

[0018] Source: Spatial Euclidean distance formula, see Chapter 7 of "Advanced Mathematics (Tongji 6th Edition)";

[0019] Technical benefits: It can determine in real time whether the robot has deviated from the center of the trajectory, thus improving the accuracy of subsequent scoring.

[0020] Preferably, the specific steps for calculating the score in S2 are as follows:

[0021] After obtaining the state parameter vector, the S21 system performs normalization preprocessing on each input data, including unit unification, dynamic range compression and outlier removal for tension change, tilt angle, lateral acceleration and centroid offset distance, to ensure that the contribution ratio of each parameter in the scoring function is controlled. After processing, a standardized vector is formed, which is convenient for subsequent function structure stability calculation and can be adapted to different types of operating conditions.

[0022] S22 inputs standardized state parameters into the disturbance factor aggregation scoring function. Based on the preset disturbance weight coefficient, it aggregates the importance of each disturbance factor by weight and outputs a stability score value S. This score value S changes dynamically within a set range and is used to quantify the stability of the robot's current operating state. The higher the score, the more unstable the operation, providing a data basis for subsequent risk assessment and control response.

[0023] The aggregate scoring function for disturbance factors is:

[0024] ;

[0025] In the formula, Operational stability score (dimensionless);

[0026] Cable tension change (N);

[0027] Robot tilt angle (°);

[0028] Lateral acceleration (m / s²);

[0029] Distance offset (m) between the centroid and the cable axis;

[0030] : Disturbance factor weighting coefficient (0~1, satisfying );

[0031] Technical source: Custom weighting function; commonly used for normalized scoring of multi-source disturbances (see weighted average model);

[0032] Explanation of sufficiency of disclosure: Each parameter comes from the S1 acquisition results, the weights are publicly adjustable, and the evaluation score is in linear form, which facilitates simulation and optimization;

[0033] Technical effect: Quantifying the comprehensive impact of different disturbance factors on stability, providing a quantitative basis for response control.

[0034] Preferably, the specific steps for determining the risk and outputting the control command in step S3 are as follows:

[0035] The S31 system compares the disturbance score S with the judgment threshold T, which is dynamically calculated through a sliding window, in real time. The score S reflects the current level of disturbance during operation, while the judgment threshold T forms a reference range based on the historical stable operating state. When the score S exceeds the judgment threshold T, it indicates that the robot's current stability is insufficient and there is a risk of derailment or falling. The system immediately triggers the control response logic and sends control commands to each actuator.

[0036] When the condition S>T is met, the system synchronously sends control commands to three types of execution structures, which are then linked and adjusted in the following order: Clamping enhancement action: After receiving the command, the servo control motor of the clamping device increases the output torque, and the clamping force is increased through the screw mechanism, which enhances the normal friction between the robot and the track surface and improves the adhesion stability; Support angle adjustment: After receiving the adjustment command, the variable angle support mechanism automatically adjusts the support arm angle according to the current tilt angle and terrain recognition results, so that the normal direction of the contact surface is aligned with the direction of gravity as much as possible, thereby improving the support rigidity; Tension relief release: The damping regulator in the tension relief structure controls the change of the release arm angle, realizing the dispersion and absorption of some lateral impact loads, slowing down the disturbance transmission path, and effectively suppressing slippage or structural vibration; All three actions are adjusted in real time based on the response information of the feedback channel to ensure that the robot still has good track adhesion and disturbance resistance stability under complex working conditions.

[0037] Preferably, the anti-detachment system for the cable-mounted coating robot is based on the above method, and the system includes:

[0038] Status acquisition device: includes tension sensor, attitude angle sensor and triaxial accelerometer, used to collect cable tension change, tilt angle, lateral acceleration and centroid offset, and construct status parameter vector;

[0039] Scoring calculator: Used to receive a state parameter vector and calculate the score value S based on the aggregated scoring function according to the perturbation factor;

[0040] Judgment Controller: Used to receive the score value S and compare it with the sliding window dynamic threshold T to determine the risk level of the current state and generate corresponding control instructions;

[0041] The actuator assembly includes a servo clamping mechanism, an adjustable angle support structure, and a tension release mechanism, which are used to perform clamping force adjustment, support angle adjustment, and tension release operations in response to control commands, respectively.

[0042] Feedback adjustment processor: used to collect motion state feedback values ​​of the above-mentioned actuators and adjust the perturbation factor weight coefficients according to the perturbation weight self-learning function, thereby optimizing the adaptability of the scoring function.

[0043] Preferably, the status acquisition device includes:

[0044] (1) The state acquisition device consists of a tension sensor, an attitude angle sensor and a three-axis accelerometer, which are fixed to key structural parts of the robot respectively: the tension sensor is installed at the interface between the support structure and the cable to measure the real-time force value of the cable, and the output unit is Newton (N); the attitude angle sensor is installed in the area near the geometric center of gravity of the robot body to monitor its pitch and roll angles relative to the horizontal plane, and the unit is angle (°); the three-axis accelerometer is set at the connection between the drive chassis and the main support frame to capture the acceleration values ​​in the X / Y / Z directions (unit is m / s²) to reflect the lateral disturbance, longitudinal impact and vertical displacement trend. Each sensor is connected to the main control board of the system to form a stable sensing data link.

[0045] (2) All the above-mentioned sensors are started simultaneously at a fixed sampling frequency (e.g., 100Hz). The collected data are processed by the timestamp synchronization module to form a consistent data stream. The tension change value (ΔT), attitude tilt angle (θ), three-axis acceleration vector (ax, ay, az) obtained by the system, and the relative offset distance of the centroid (Δcg, in meters) derived by the centroid estimation model are uniformly converted in terms of dimensions, normalized in terms of range and aligned in terms of time sequence to form a state parameter vector: V = [ΔT, θ, ax, Δcg]. This vector is used as the input variable of the disturbance factor aggregation scoring function to comprehensively reflect the force state, tilt trend, vibration disturbance and stability offset of the robot under the current track conditions, and support the subsequent scoring, judgment and dynamic response mechanism.

[0046] Preferably, the rating calculator includes:

[0047] (1) The scoring calculator receives the complete parameter vector V=[ΔT,θ,ax, ...] constructed from the state acquisition device. After cg], the physical quantities are first normalized to units, and each term is mapped to the [0,1] interval using a linear normalization method:

[0048] ;

[0049] In the formula, For the original physical quantity;

[0050] and These are the historical minimum and maximum values ​​of the variable, respectively. After normalization, the calculator performs sliding window mean filtering (window width of 5-10 frames) on various types of data, and supplements it with median filtering algorithm to filter out short-term mutation interference, improve the stability of the input data response to the scoring model, and ensure that the subsequent function structure has repeatability and convergence.

[0051] (2) The scoring calculator has an embedded perturbation factor aggregation scoring function, which adopts a weighted linear combination model.

[0052] ;

[0053] In the formula, , , , The normalized tension disturbance, attitude tilt angle, absolute value of lateral acceleration, and center of mass offset distance; The perturbation factor weighting coefficient satisfies Its value can be set through simulation training or empirical evaluation, and the output score is determined. [0,1], where: This indicates stable operation; This indicates a slight disturbance; This indicates a high level of stability risk; the scoring result will serve as a key basis for the subsequent controller to determine whether to trigger the stability response mechanism, and has clear mathematical support and physical significance.

[0054] Preferably, the determination controller includes:

[0055] (1) The controller receives the score value S output by the score calculator and updates the judgment threshold T in real time from the historical score data through the sliding window mechanism. The threshold T is calculated based on the mean and standard deviation of the recent N score values. It has adaptive capability and can be dynamically adjusted according to the degree of disturbance of the operating environment, so as to ensure that the risk identification is real-time and accurate, and avoid the problem of misjudgment or delayed judgment caused by fixed threshold.

[0056] The principle expression of the sliding window mechanism is:

[0057] ;

[0058] In the formula, T: the current scoring threshold;

[0059] : The mean of the rating sequence in period t;

[0060] : The standard deviation of the scoring sequence in period t;

[0061] k: Risk adjustment factor (recommended value: 1.5~2.5);

[0062] The time window length N is a fixed preset value (e.g., N = number of samples within 20 seconds);

[0063] Technical source: Slippage anomaly detection criteria commonly used in statistics; applicable to dynamic risk prediction;

[0064] Statement of sufficiency of disclosure: The algorithm structure is completely public, the parameters are configurable, and historical scoring samples can be stored in the system in real time;

[0065] Technical effect: The system can adapt to different disturbance environments and dynamically adjust the judgment sensitivity to avoid hypersensitive or delayed triggering;

[0066] (2) When the score value S is greater than the current judgment threshold T, the judgment controller judges that the robot's operating state is unstable and generates a corresponding level of control response instruction accordingly. The risk level is divided into multiple levels, each level corresponds to a control strategy of different strengths. The control instruction will be transmitted synchronously to the execution device group to trigger clamping, support and tension adjustment response actions to achieve accurate stability recovery and risk avoidance of fall.

[0067] Preferably, the actuator group:

[0068] (1) The servo gripping mechanism in the execution device group includes a pair of adjustable mechanical grippers, which are driven and controlled by a servo motor. The gripping force is adjusted by a screw mechanism. After receiving the clamping enhancement command issued by the judgment controller, the gripping mechanism automatically increases the clamping force to enhance the frictional contact between the robot body and the track or support surface, prevent loosening or slippage caused by sudden tension changes or inertial offset, and ensure operational stability.

[0069] The principle expression for clamping force control output is:

[0070] ;

[0071] In the formula, : Current clamping force output of the clamping device (N);

[0072] Current rating;

[0073] : Current threshold;

[0074] Proportional gain (unit: N);

[0075] Differential gain (unit: N·s);

[0076] : Rate of change of the score-threshold difference (speed of risk change);

[0077] Technical source: PD controller structure in automatic control theory, see Hu Shousong's textbook "Principles of Automatic Control" for details;

[0078] Explanation of sufficiency of disclosure: The parameter structure is adjustable and the model is simulable, clearly reflecting the control path of "high score → increased clamping force";

[0079] Technical effect: Achieves rigid-flexible linkage anti-detachment control; the more drastic the score change, the faster the gripper responds, thus improving system stability.

[0080] (2) The adjustable angle support structure adopts a multi-link telescopic mechanism to adjust the angle of the support surface, which can be synchronously corrected according to the current terrain changes or attitude inclination; the tension relief mechanism is equipped with an angle control arm and a damping adjuster, which can change the release angle and relief resistance according to the control command, so as to absorb part of the impact load when excessive lateral disturbance occurs, and avoid direct transmission to the gripper area. The three links respond in a coordinated manner to form a three-dimensional stability protection system.

[0081] Preferably, the feedback adjustment processor:

[0082] (1) The feedback adjustment processor is connected to the control terminal interface of the servo clamping structure, support structure and tension relief mechanism. After the control command is executed, the system synchronously collects the action feedback parameters, including the clamping force change curve. , offset angle of supporting structure and tension release angular velocity Each feedback data point is normalized using a sliding window and compared with the target setpoint to form a sequence of execution deviation values.

[0083] ;

[0084] In the formula, Indicates the first A feedback variable (such as clamping force) is used as the feedback error input, which will be used in the control-response effect analysis and scoring deviation correction path.

[0085] (2) After collecting the feedback of the execution results, the feedback adjustment processor calls the perturbation weight self-learning function. Based on the correlation between the scoring deviation and the response efficiency, the weight coefficients of each perturbation factor in the scoring function are dynamically updated. The function adopts the exponential decay and sliding fitting strategy to automatically strengthen or weaken the contribution ratio of each factor according to the recent multiple response results, thereby improving the adaptability of the scoring function to the operating state and the accuracy of actual risk judgment.

[0086] The perturbation weight self-learning function is:

[0087] ;

[0088] In the formula, The weight of the i-th perturbation factor in the next cycle;

[0089] The absolute value of the current i-th disturbance factor;

[0090] Learning rate (usually taken as ) ;

[0091] , : Normalization of the total amount of disturbance factor;

[0092] Explanation of sufficiency: The computational logic is clear, the input variables are well-defined, and the boundaries are controllable;

[0093] Technical effects: Enhances the self-learning ability of the scoring system and improves the sensitivity and adaptability of the scoring function to the actual perturbation distribution.

[0094] The beneficial effects of this invention are as follows:

[0095] 1. This invention acquires multi-source disturbance data in real time by deploying tension sensors, attitude angle sensors, and triaxial acceleration sensors on key parts of the robot body, and constructs a structured state parameter vector, thereby achieving a comprehensive quantitative characterization of the operating state. Compared with the traditional method of judging risk based on a single threshold, this multi-dimensional information fusion mechanism can simultaneously perceive force, attitude, and dynamic disturbances, effectively improving the ability to identify and judge potential detachment risks. It is particularly suitable for dynamic stability monitoring in high-altitude flexible track scenarios, ensuring safe and continuous operation of the robot during long-distance operation.

[0096] 2. This invention uses a perturbation factor aggregation scoring function to weight the state vector and introduces a sliding window dynamic threshold construction mechanism to make the scoring judgment real-time, flexible and adaptable to working conditions. Compared with the fixed threshold judgment strategy, this method can dynamically adjust the judgment criteria in scenarios with severe perturbation or complex trajectory changes, realize early identification and graded response to unstable working conditions. The scoring result directly drives the control module to output adjustment commands, quickly start the clamping force enhancement, support angle adjustment and tension relief linkage mechanism, and intervene immediately when the system detects the critical stability boundary, effectively improving the robot's attachment stability and anti-detachment ability.

[0097] 3. This invention constructs a complete action-feedback closed-loop mechanism. By adjusting the processor to collect motion feedback parameters such as clamping force, support angle, and tension release of the execution device, and calling the perturbation weight self-learning function to dynamically optimize the parameters in the scoring function, this adaptive mechanism enables the system to automatically adjust the contribution ratio of perturbation factors based on historical response effects. This avoids the problem of misjudgment or slow response caused by fixed models in complex environments, and realizes the stable operation of the robot under multiple perturbations and the ability to suppress risk through self-learning. Attached Figure Description

[0098] Figure 1 This is a flowchart of the anti-detachment method for the cable coating robot of the present invention;

[0099] Figure 2 This is a flowchart of the anti-detachment system for the cable coating robot of the present invention. Detailed Implementation

[0100] 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.

[0101] like Figures 1 to 2 As shown, this embodiment of the invention provides a method for preventing the cable coating robot from detaching. The specific steps of this method are as follows:

[0102] S1 acquires state parameters: By using tension sensors, attitude angle sensors and three-axis accelerometers set on the robot body, the changes in cable tension, tilt angle, lateral acceleration and centroid offset distance are acquired in real time to construct a set of state parameters;

[0103] S2 calculates the score: The state parameter vector is input into the disturbance factor aggregation score function to obtain the score S reflecting the current operational stability of the robot;

[0104] S3 determines the risk and outputs control commands: It compares the score value S with the judgment threshold T dynamically calculated based on the sliding window. If the score value S is greater than the judgment threshold T, it initiates the corresponding control command and drives the execution device to complete the process in sequence; it increases the clamping force of the clamping device; and it adjusts the contact angle of the support structure.

[0105] Adjust the release angle of the tension relief device to suppress the risk of detachment and maintain trajectory adhesion stability.

[0106] Example 1: Implementation plan for a scoring-driven method to prevent cable detachment in a cable-operated vehicle.

[0107] When applying a coating to the surface of the main cable of a high-altitude bridge, a cable-mounted robot with four-wheel drive and a biomimetic gripping structure is used. Its walking path runs along a steel cable with a diameter of about 80mm. Considering the actual situation of wind load, gravity offset and local cable vibration interference, this embodiment introduces a multi-source perception and disturbance scoring collaborative control mechanism.

[0108] Tension sensors (model FUTEK LSB200, range 0–100N) are installed at the connection points at both ends of the vehicle body, and an IMU module (integrating a triaxial acceleration and attitude angle sensor, such as Bosch BNO055) is set in the center to collect tension changes, lateral acceleration (±16m / s²), pitch / roll angle (±180°) and center of gravity offset trend in real time.

[0109] The system samples and forms a state parameter vector containing four types of disturbance factors every second. After normalization, the vector is input into the disturbance aggregation scoring function. The scoring function outputs a stability score value S, with the dynamic range defined as [0,1]. Where S>0.65 is considered a high-risk area. The sliding window length is set to 60 frames (about 1 minute). The historical mean μ and standard deviation σ are calculated to form a dynamic threshold T=μ+1.5σ.

[0110] When the score S exceeds T, a control response is immediately triggered. The clamping mechanism (servo-driven electric cylinder) responds to the signal, increasing the clamping pressure from 50N to 80N; the support structure (aluminum alloy elastic legs) is servo-driven to adjust the contact angle by ±3°; and the tension relief mechanism (spring-linked slide) increases the release angle to 18°. All three actions are synchronized within 0.8 seconds.

[0111] The feedback adjustment processor collects clamping pressure values ​​(via a force-sensitive resistor), support angle displacement feedback (via a rotary encoder), and tension release response delay. The system then calls a self-learning function to fine-tune the scoring function parameters. This solution successfully avoided three derailment attempts caused by lateral wind pressure during field testing, effectively improving the safety of high-altitude operations and the stability of the robot's track.

[0112] In S1, the acquisition of state parameters refers to the deployment of tension sensors, attitude angle sensors, and triaxial accelerometers at key operating points of the cable coating robot body. The tension sensors are installed between the main load-bearing structures on the robot's load side to continuously monitor changes in cable stress during operation and output cable tension values ​​(in N). The attitude angle sensors are located at the geometric center of the robot body to obtain the tilt angle (in °) of the coating module relative to the horizontal reference plane in real time. The triaxial accelerometers are installed at the connection between the chassis and the active drive component to collect X / Y / Z acceleration changes (in m / s²) to assess the degree of trajectory disturbance and the trend of operational inertia. The aforementioned sensors operate synchronously at a sampling frequency of 100Hz, outputting standardized physical quantity streams as basic data channels connected to the state modeling module. Based on data collected by various sensors, the system synchronously collects data including: real-time changes in cable tension (in Newtons), robot tilt angle (in degrees), and lateral acceleration (in m / s²). This data is further combined with the relative offset distance of the center of mass (in meters) measured by the center of gravity position detection device. These values ​​are quantified, time-aligned, and encoded to form a state parameter vector with a fixed structure. This vector serves as the input variable for the scoring function, reflecting characteristics such as the robot's operating posture, track adhesion, and disturbance amplitude, providing a quantitative basis for subsequent risk identification and control strategy generation.

[0113] The S2 calculation score refers to the system's normalization preprocessing of various input data after acquiring the state parameter vector. This includes standardizing units, compressing dynamic range, and removing outliers for tension changes, tilt angles, lateral acceleration, and centroid offset distance. This ensures that the contribution ratio of each parameter in the scoring function is controlled, resulting in a standardized vector that facilitates subsequent function structure stability calculations and can adapt to different types of operating conditions. The standardized state parameters are then input into the disturbance factor aggregation scoring function. Based on preset disturbance weight coefficients, the importance of each disturbance factor is weighted and aggregated to output a stability score S. This score S dynamically changes within a set range to quantify the stability of the robot's current operating state. A higher score indicates less stable operation, providing a data basis for subsequent risk assessment and control response.

[0114] In S3, risk assessment and control command output refer to the system comparing the disturbance score S with the judgment threshold T dynamically calculated through a sliding window in real time. The score S reflects the current level of operational disturbance, while the judgment threshold T forms a reference range based on historical stable operating states. When the score S exceeds the judgment threshold T, it indicates that the robot's current stability is insufficient, posing a risk of derailment or fall. The system immediately triggers the control response logic, issuing adjustment commands to each actuator. When the condition S>T is met, the system synchronously sends control commands to the three types of actuators, adjusting them in the following order: Clamping enhancement action: After receiving the command, the servo control motor of the clamping device increases its output torque, through... The screw mechanism increases clamping force, enhances the normal friction between the robot and the track surface, and improves adhesion stability. Support angle adjustment: After receiving an adjustment command, the variable angle support mechanism automatically adjusts the support arm angle based on the current tilt angle and terrain recognition results, aligning the contact surface normal direction as closely as possible with the direction of gravity, thus improving support rigidity. Tension relief release: The damping regulator in the tension relief structure controls the change in the release arm angle, achieving partial dispersion and absorption of lateral impact loads, slowing down the disturbance transmission path, and effectively suppressing slippage or structural vibration. All three actions are adjusted in real-time closed-loop based on the response information of the feedback channel to ensure that the robot still has good track adhesion and disturbance resistance stability under complex working conditions.

[0115] This embodiment provides an anti-detachment system for the cable coating robot, and the system includes:

[0116] Status acquisition device: includes tension sensor, attitude angle sensor and triaxial accelerometer, used to collect cable tension change, tilt angle, lateral acceleration and centroid offset, and construct status parameter vector;

[0117] Scoring calculator: Used to receive a state parameter vector and calculate the score value S based on the aggregated scoring function according to the perturbation factor;

[0118] Judgment Controller: Used to receive the score value S and compare it with the sliding window dynamic threshold T to determine the risk level of the current state and generate corresponding control instructions;

[0119] The actuator assembly includes a servo clamping mechanism, an adjustable angle support structure, and a tension release mechanism, which are used to perform clamping force adjustment, support angle adjustment, and tension release operations in response to control commands, respectively.

[0120] Feedback adjustment processor: used to collect motion state feedback values ​​of the above-mentioned actuators and adjust the perturbation factor weight coefficients according to the perturbation weight self-learning function, thereby optimizing the adaptability of the scoring function.

[0121] The state acquisition device consists of a tension sensor, an attitude angle sensor, and a three-axis accelerometer, which are fixed to key structural parts of the robot: the tension sensor is located at the interface between the support structure and the cable to measure the real-time force on the cable, with the output unit being Newtons (N); the attitude angle sensor is installed near the geometric center of gravity of the robot body to monitor its pitch and roll angles relative to the horizontal plane, with the unit being degrees (°); and the three-axis accelerometer is located at the connection between the drive chassis and the main support frame to capture X / Y / Z axes. Acceleration values ​​in three directions (in m / s²) are used to reflect the trends of lateral disturbance, longitudinal impact, and vertical displacement. Each sensor is connected to the main control board of the system to form a stable sensing data link. All the above-mentioned sensors are started simultaneously at a fixed sampling frequency (e.g., 100Hz). The collected data is processed by the timestamp synchronization module to form a consistent data stream. The tension change value (ΔT), attitude tilt angle (θ), three-axis acceleration vector (ax, ay, az), and the relative offset distance of the centroid (Δcg, in meters) obtained by the centroid estimation model are used to perform unified dimension conversion, range normalization, and time alignment of various physical quantities to form a state parameter vector: V = [ΔT, θ, ax, Δcg]. This vector serves as the input variable of the disturbance factor aggregation scoring function, comprehensively reflecting the robot's force state, tilt trend, vibration disturbance, and stability offset under the current track conditions, supporting subsequent scoring, judgment, and dynamic response mechanisms.

[0122] Among them, the scoring calculator refers to the calculator that receives the complete parameter vector V=[ΔT,θ,ax, ...] constructed from the state acquisition device. After cg], the physical quantities are first normalized to units, and each term is mapped to the [0,1] interval using a linear normalization method:

[0123] ;

[0124] In the formula, For the original physical quantity, and These are the historical minimum and maximum values ​​of the variable, respectively. After normalization, the calculator performs sliding window mean filtering (window width 5-10 frames) on various types of data, supplemented by median filtering to filter out short-term abrupt changes and improve the stability of the input data's response to the scoring model, ensuring the repeatability and convergence of the subsequent function structure. The scoring calculator has an embedded perturbation factor aggregation scoring function, which adopts a weighted linear combination model.

[0125] ;

[0126] In the formula, , , , The normalized tension disturbance, attitude tilt angle, absolute value of lateral acceleration, and center of mass offset distance; The perturbation factor weighting coefficient satisfies Its value can be set through simulation training or empirical evaluation, and the output score is determined. [0,1], where: This indicates stable operation; This indicates a slight disturbance; This indicates a high level of stability risk; the scoring result will serve as a key basis for the subsequent controller to determine whether to trigger the stability response mechanism, and has clear mathematical support and physical significance.

[0127] The judgment controller receives the score value S output by the scoring calculator and updates the judgment threshold T in real time from historical scoring data through a sliding window mechanism. This threshold T is calculated based on the mean and standard deviation of the last N scores and has adaptive capabilities, dynamically adjusting to changes in the degree of disturbance in the operating environment. This ensures that risk identification is real-time and accurate, avoiding misjudgments or delayed judgments caused by fixed thresholds. When the score value S is greater than the current judgment threshold T, the judgment controller determines that the robot's operating state is unstable and generates corresponding control response instructions. The risk level is divided into multiple levels, each corresponding to a control strategy of different strengths. The control instructions are synchronously transmitted to the actuator group to trigger clamping, support, and tension adjustment response actions to achieve precise stability restoration and avoidance of detachment risks.

[0128] The actuator group refers to the servo gripping mechanism, which includes a pair of adjustable mechanical grippers driven and controlled by a servo motor. The gripping force is adjusted via a screw mechanism. Upon receiving a clamping enhancement command from the judgment controller, the gripping mechanism automatically increases the clamping force to enhance the frictional contact between the robot body and the track or support surface, preventing loosening or slippage caused by sudden tension changes or inertial shifts, thus ensuring operational stability. The adjustable angle support structure uses a multi-link telescopic mechanism to adjust the angle of the support surface, which can be synchronously corrected according to changes in terrain or tilt. The tension relief mechanism is equipped with an angle control arm and a damping adjuster, which can change the release angle and relief resistance according to control commands to absorb some of the impact load when excessive lateral disturbances occur, preventing direct transmission to the gripper area. The three components work together to form a three-dimensional stability protection system.

[0129] The feedback adjustment processor is connected to the motion control port of the servo clamping mechanism, support structure, and tension relief mechanism. It collects the action execution values ​​and transmission feedback signals during the response process, including parameters such as clamping force change curves, support angle adjustment displacement, and tension release angular velocity. The system performs normalization processing and error detection on the feedback data to form control command-action result comparison data, providing input basis for subsequent adaptive optimization of disturbance factor weights. After collecting the feedback of the action execution results, the feedback adjustment processor calls the disturbance weight self-learning function. Based on the correlation between scoring deviation and response efficiency, it dynamically updates the weight coefficients of each disturbance factor in the scoring function. This function adopts an exponential decay and sliding fitting strategy, automatically strengthening or weakening the contribution ratio of each factor based on recent multiple response results, thereby improving the adaptability of the scoring function to the operating state and the accuracy of actual risk judgment.

[0130] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0131] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for preventing the coating of cables on a cable-carrying robot from falling off, characterized in that: The specific steps of the method for preventing the cable roll coating robot from falling off are as follows: S1 acquires state parameters: By using tension sensors, attitude angle sensors and three-axis accelerometers set on the robot body, the changes in cable tension, tilt angle, lateral acceleration and centroid offset distance are acquired in real time to construct a set of state parameters; S2 calculates the score: The state parameter vector is input into the disturbance factor aggregation score function to obtain the score S reflecting the current operational stability of the robot; S3 determines the risk and outputs control commands: It compares the score value S with the judgment threshold T obtained by dynamic calculation based on the sliding window. When the score value S is greater than the judgment threshold T, it initiates the corresponding control command and drives the execution device to complete the following in sequence: increase the clamping force of the clamping device, adjust the contact angle of the support structure, and adjust the release angle of the tension relief device; in order to suppress the risk of falling off and maintain the stability of the trajectory adhesion.

2. The method for preventing the coating of cables on a cable-carrying robot according to claim 1, characterized in that: The specific steps for collecting state parameters in S1 are as follows: The S11 has tension sensors, attitude angle sensors and three-axis acceleration sensors installed at key structural points of the coating robot body in the cable-line model. The tension sensors are located between the load-bearing structures; the attitude angle sensors are installed at the center of the robot body to obtain the tilt angle between the coating unit and the support structure on the running trajectory; and the three-axis acceleration sensors are fixed at the connection between the chassis and the drive module. Based on data collected by various sensors, the S12 system synchronously collects data including: real-time changes in cable tension, robot tilt angle, and lateral acceleration. It further combines the relative offset distance of the center of mass measured by the center of mass position detection device, quantizes, aligns and encodes the above values, and forms a state parameter vector with a fixed structure.

3. The method for preventing the coating of cables on a cable-carrying robot according to claim 2, characterized in that: The specific steps for calculating the score value in S2 are as follows: After obtaining the state parameter vector, S21 performs normalization preprocessing on each input data, including unit unification, dynamic range compression and outlier removal for tension change, tilt angle, lateral acceleration and centroid offset distance, to ensure that the contribution ratio of each parameter in the scoring function is controlled and a standardized vector is formed after processing. S22 inputs standardized state parameters into the disturbance factor aggregation scoring function. Based on the preset disturbance weight coefficient, it aggregates the importance of each disturbance factor by weight and outputs a stability score value S. This score value S changes dynamically within a set range and is used to quantify the stability of the robot's current operating state. The higher the score, the more unstable the operation.

4. The method for preventing the coating of cables on a cable-carrying robot according to claim 3, characterized in that: The specific steps for determining the risk and outputting control commands in S3 are as follows: The S31 system compares the disturbance score S with the judgment threshold T, which is dynamically calculated through a sliding window, in real time. The score S reflects the current level of disturbance during operation, while the judgment threshold T forms a reference range based on the historical stable operating state. When the score S exceeds the judgment threshold T, it indicates that the robot's current stability is insufficient and there is a risk of derailment or falling. The system immediately triggers the control response logic and sends control commands to each actuator. The S32 control response includes three linked adjustment actions: first, increasing the clamping force of the clamping device to improve adhesion stability by enhancing contact pressure; second, adjusting the contact angle of the support structure to adapt to terrain deviations or attitude tilts; and third, adjusting the release angle of the tension relief device to reduce the transmission of lateral disturbances.

5. A cable coating robot anti-detachment system, characterized in that: A method for preventing cable coating robots according to any one of claims 1 to 4, wherein the system comprises: Status acquisition device: includes tension sensor, attitude angle sensor and triaxial accelerometer, used to collect cable tension change, tilt angle, lateral acceleration and centroid offset, and construct status parameter vector; Scoring calculator: Used to receive a state parameter vector and calculate the score value S based on the aggregated scoring function according to the perturbation factor; Judgment Controller: Used to receive the score value S and compare it with the sliding window dynamic threshold T to determine the risk level of the current state and generate corresponding control instructions; The actuator assembly includes a servo clamping mechanism, an adjustable angle support structure, and a tension release mechanism, which are used to perform clamping force adjustment, support angle adjustment, and tension release operations in response to control commands, respectively. Feedback adjustment processor: used to collect motion state feedback values ​​of the above-mentioned actuators and adjust the perturbation factor weight coefficients according to the perturbation weight self-learning function, thereby optimizing the adaptability of the scoring function.

6. The anti-detachment system for a cable-mounted coating robot according to claim 5, characterized in that: The status acquisition device includes: (1) The state acquisition device consists of a tension sensor, an attitude angle sensor and a three-axis accelerometer. The tension sensor is installed at the connection between the support mechanism and the cable to monitor the force state of the cable in real time during operation. The attitude angle sensor is installed near the center of gravity of the robot to obtain the tilt angle of the robot body. The three-axis accelerometer is deployed in the drive unit module to capture lateral disturbances, longitudinal impacts and vertical displacement trends. (2) The above-mentioned sensors collect data at a fixed sampling frequency and build a consistent data channel through a synchronous acquisition mechanism. The system encodes and quantizes the collected tension change value, tilt angle information, triaxial acceleration vector and centroid offset derived from inertia into a standard format state parameter vector.

7. The anti-detachment system for a cable-mounted coating robot according to claim 6, characterized in that: The scoring calculator includes: (1) After receiving the complete parameter vector from the state acquisition device, the scoring calculator first performs unit normalization and time sequence alignment on various physical quantities to ensure that the tension value, tilt angle, acceleration and centroid offset are comparable on the same scale. Then, the data is preprocessed using dynamic amplitude limiting and noise filtering algorithms to remove short-term mutations and external interference, so that the input data meets the structural stability requirements of the scoring model. (2) The scoring calculator has a built-in perturbation factor aggregation scoring function. Based on the set perturbation weight coefficient, it performs weighted superposition calculation of standardized tension perturbation, attitude deviation and acceleration mutation factor, and outputs a score value S representing the current operational stability.

8. The anti-detachment system for a cable coating robot according to claim 7, characterized in that: The determination controller includes: (1) The decision controller receives the score value S output by the score calculator and updates the decision threshold T in real time from the historical score data through the sliding window mechanism. The threshold T is calculated based on the mean and standard deviation of the recent N score values. (2) When the score value S is greater than the current judgment threshold T, the judgment controller judges that the robot's operating state is unstable and generates a corresponding level of control response instruction accordingly. The risk level is divided into multiple levels, each level corresponds to a control strategy of different strengths, and the control instruction will be synchronously transmitted to the execution device group to trigger clamping, support and tension adjustment response actions.

9. The anti-detachment system for a cable-mounted coating robot according to claim 8, characterized in that: The actuator group includes: (1) The servo gripping mechanism in the execution device group includes a pair of adjustable mechanical grippers, which are driven and controlled by a servo motor. The gripping force is adjusted by a screw mechanism. After receiving the clamping enhancement command issued by the judgment controller, the gripping mechanism automatically increases the clamping force to enhance the frictional contact between the robot body and the track or support surface. (2) The adjustable angle support structure adopts a multi-link telescopic mechanism to realize the angle adjustment of the support surface. It can be synchronously corrected according to the current terrain changes or attitude inclination. The tension relief mechanism is equipped with an angle control arm and a damping adjuster, which can change the release angle and relief resistance according to the control command.

10. The anti-detachment system for a cable-mounted coating robot according to claim 9, characterized in that: The feedback regulation processor includes: (1) The feedback adjustment processor is connected to the motion control port of the servo clamping mechanism, the support structure and the tension relief mechanism to collect the action execution value and transmission feedback signal during the response process, including the clamping force change curve, the support angle adjustment displacement and the tension release angular velocity parameter; (2) After collecting the feedback of the execution results, the feedback adjustment processor calls the perturbation weight self-learning function and dynamically updates the weight coefficients of each perturbation factor in the scoring function based on the correlation between the scoring deviation and the response efficiency.

Images