An aerial maintenance robot's working arm motion trajectory optimization control method

Abstract

This application relates to the field of intelligent manufacturing equipment technology, specifically to a method for optimizing the motion trajectory control of the working arm of an overhead maintenance robot. The method includes: determining the state parameters of the overhead maintenance robot during the current acquisition cycle; determining the difference in guide wire distance between the front / rear traveling wheels between the current acquisition cycle and the next acquisition cycle; determining the force interference degree of any traveling wheel in the current acquisition cycle; adjusting a fixed constant of the sliding mode control convergence rate for the next acquisition cycle based on the guide wire distance difference and the force interference degree; and controlling the opening degree of the traveling arm in the next acquisition cycle using a sliding mode control algorithm based on the adjusted fixed constant. This application aims to achieve precise optimization control of the working arm's motion trajectory, improving the robot's walking stability and maintenance safety.

Description

Technical Field

[0001] This application relates to the field of intelligent manufacturing equipment technology, specifically to a method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot. Background Technology

[0002] Overhead transmission lines in power distribution networks, after long-term operation, are often subject to increased aging and fault rates due to external factors such as wind, rain, and impacts from foreign objects. As the power grid expands, transmission lines become crucial for the stable operation of the power system, making regular maintenance essential. Traditional methods combining manual labor and drones are typically inefficient and pose certain safety risks. With the development of intelligent robotics, overhead maintenance robots are widely used in overhead line maintenance. The efficiency of these robots is closely related to the stability of their arm movements. Therefore, in the intelligent manufacturing equipment industry for overhead maintenance robots, researching a precise method for optimizing the arm's motion trajectory is of great significance for improving the efficiency and safety of overhead maintenance robots.

[0003] In the motion trajectory control method of the working arm (such as the walking arm) of the overhead maintenance robot, the opening of the working arm is mainly adjusted by controlling the extension and contraction of the pneumatic spring, thereby optimizing the motion trajectory of the working arm and ensuring that the walking wheels of the overhead maintenance robot move stably along the overhead line. The existing technology mainly uses the sliding mode control algorithm to optimize the motion trajectory control of the working arm of the overhead maintenance robot. To a certain extent, it can maintain the stable forward movement of the overhead maintenance robot. However, the approach rate of the sliding mode control mainly depends on the sign function and the sliding surface, which makes it difficult to optimize chattering and convergence time at the same time. During the forward movement of the overhead maintenance robot, it is usually affected by unknown external factors, such as strong winds at high altitudes causing the overhead line to sway, affecting the motion stability of the overhead maintenance robot. The accuracy of its motion trajectory control mainly depends on the approach rate. If the approach rate is not set reasonably, the opening of the walking arm will be difficult to adapt to the dynamically swaying overhead line, resulting in severe shaking or even derailment of the overhead maintenance robot, making it difficult to efficiently complete the overhead line maintenance task. Summary of the Invention

[0004] To address the aforementioned technical problems, this application provides an embodiment of a method for optimizing the motion trajectory of the working arm of an overhead maintenance robot, employing the following technical solution: This application provides a method for optimizing the motion trajectory control of the working arm of an overhead maintenance robot. The overhead maintenance robot includes at least four independent walking arms, which are evenly installed on both sides of the robot's body. The distance between the front and rear walking wheels, which are elastically hinged to the walking arms and located on both sides of the overhead maintenance robot, is adaptively controlled using a sliding mode control algorithm. The method includes the following steps: Determine the state parameters of the overhead maintenance robot during the current data acquisition cycle. The state parameters include the distance between overhead conductors, the vector force on the walking wheels, and the opening of the walking arm. Predict the distance between the overhead conductors where the front / rear traveling wheels are located in the next acquisition cycle; determine the conductor distance difference between the front / rear traveling wheels and the next acquisition cycle based on the distance prediction value. The conductor distance difference is used to measure the degree of fluctuation difference of the overhead conductors due to external factors between adjacent acquisition cycles. The force interference degree of any walking wheel in the current acquisition cycle is determined based on the vector force. The force interference degree is used to measure the degree of force fluctuation of the walking wheel due to external factors. The fixed constant of the sliding mode control approach rate for the next acquisition cycle is adjusted based on the difference in conductor distance and the degree of force interference. Based on the adjusted fixed constant, the opening of the traveling arm in the next acquisition cycle is controlled by the sliding mode control algorithm.

[0005] Preferably, determining the distance between the overhead conductors includes: Images of the traveling wheel moving along the overhead conductor are captured using a binocular depth camera; The wheel-line contact area between the walking wheel and the overhead conductor in the image is identified using a pre-trained target detection model; The three-dimensional center coordinates of the wheel-line contact area are obtained using binocular stereo matching technology; The distance between the overhead conductors is calculated based on the three-dimensional center coordinates.

[0006] Preferably, the predicted distance between the overhead conductors where the front / rear wheels are located in the next acquisition cycle includes: The distances between the front and rear walking wheels of the overhead conductors collected during all historical data collection periods during the operation of the overhead maintenance robot are obtained and arranged into a conductor distance sequence according to the collection time. The conductor distance sequence is used as input to the ARIMA prediction algorithm to obtain the predicted distance between the overhead conductors where the front and rear traveling wheels are located in the next acquisition cycle.

[0007] Preferably, the determination of the conductor distance difference includes: Calculate the first-order difference sequence of the sequence composed of the predicted distances between the overhead conductors where the front and rear wheels are located in the next acquisition cycle, and obtain the average value of the absolute values ​​of all elements in the first-order difference sequence; For the front and rear wheels of the overhead maintenance robot, the difference in the guide wire distance between the front and rear wheels between the current acquisition cycle and the next acquisition cycle is determined based on the difference between the average value calculated in the current acquisition cycle and the average value calculated in the next acquisition cycle.

[0008] Preferably, the determination of the degree of force disturbance includes: Obtain the vector force experienced by any walking wheel at all sampling moments within the current acquisition cycle and assemble it into a vector force sequence in chronological order; Calculate the coefficient of variation of the magnitudes of all vector forces in the vector force sequence; Calculate the mean cosine similarity between adjacent vector forces in the vector force sequence; The force interference degree of any walking wheel in the current acquisition cycle is determined by weighted summation of the mean of the coefficient of variation and the cosine similarity.

[0009] Preferably, the fixed constant for adjusting the sliding mode control approach rate of the next acquisition cycle based on the difference in conductor distance and the degree of force interference includes: Obtain the current fixed constant of the approach rate of the sliding mode control algorithm in the current acquisition period; Calculate the mean difference in guide wire distance between the front and rear wheels between the current acquisition cycle and the next acquisition cycle; The adjustment range is determined based on the average force interference of all the wheels of the overhead maintenance robot during the current data collection cycle. Based on the current fixed constant, the mean value of the conductor distance difference, and the adjustment range, determine the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle.

[0010] Preferably, determining the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle based on the current fixed constant, the mean of the conductor distance difference, and the adjustment range includes: If the mean value of the conductor distance difference is greater than 0, the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle is reduced by the adjustment range based on the current fixed constant. If the mean value of the conductor distance difference is less than 0, the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle is increased by the adjustment range based on the current fixed constant. If the mean value of the conductor distance difference is equal to 0, then the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle remains unchanged as the current fixed constant.

[0011] Preferably, controlling the opening degree of the traveling arm in the next acquisition cycle through the sliding mode control algorithm includes: The target distance between the front and rear traveling wheels in the next acquisition cycle is determined based on the average value of the predicted distance between the overhead conductors where the front and rear traveling wheels are located in the next acquisition cycle, and the target opening width of the front and rear traveling arms in the next acquisition cycle is also determined. The target opening width is used to determine the target opening degree of any traveling arm in the next acquisition cycle; The target opening of all arms in the next acquisition cycle and the real-time opening of all arms in the current acquisition cycle are input into the sliding mode control algorithm. The sliding mode control algorithm uses an adjusted fixed constant in the next acquisition cycle and outputs a control signal to adjust the extension and retraction of the pneumatic spring.

[0012] Preferably, the value range of the fixed constant after adjustment is 0.1 to 1. If the adjusted value exceeds the value range, the fixed constant is set as the boundary value of the value range.

[0013] Preferably, the target opening width of the front / rear walking arms is half the difference between the target distance between the front / rear walking wheels and the width of the overhead maintenance robot frame.

[0014] This application has at least the following beneficial effects: 1. This application addresses the problem that the fixed constant of the approach rate in traditional sliding mode control is difficult to adapt to the dynamic swaying of overhead conductors. By predicting the conductor distance in the next acquisition cycle and combining the conductor distance difference and force interference in the current cycle, the fixed constant of the approach rate is adaptively adjusted. This enables the sliding mode control to dynamically optimize control parameters according to the real-time changes of external interference, avoids the contradiction between chattering and convergence time, and effectively improves the motion stability of the overhead maintenance robot under interference such as strong winds.

[0015] 2. This application uses the ARIMA prediction algorithm to predict the conductor distance in future acquisition cycles, and combines it with the first-order difference sequence analysis to analyze the difference in distance fluctuation between adjacent cycles. This can detect the fluctuation trend of overhead conductors under external interference in advance, provide a basis for the forward adjustment of the fixed constant of the sliding mode control approach rate, and enhance the control system's ability to predict and respond to dynamic environments.

[0016] 3. This application constructs a force interference index by calculating the mean of the coefficient of variation of the magnitude of the vector force of the walking wheel and the cosine similarity of adjacent vector forces. This index can accurately quantify the degree of force fluctuation of the walking wheel under external interference, avoiding the defect that a single-dimensional signal cannot fully reflect the influence of interference, and providing a more comprehensive decision-making basis for adjusting the constant of the convergence rate.

[0017] 4. This application determines the adjustment direction based on the positive or negative sign of the conductor distance difference and determines the adjustment range in combination with the average value of the force disturbance, so that the adjustment of the fixed constant of the sliding mode control approach rate not only conforms to the trend direction of the disturbance change, but also matches its intensity, thus realizing the precise adaptive correction of the fixed constant and ensuring that the control parameters are always in the optimal range.

[0018] 5. This application uses a sliding mode control algorithm with an adjusted fixed constant for the target opening and the real-time opening input to control the extension and retraction of the pneumatic spring and achieve precise adjustment of the traveling arm opening. Combined with the constraint of the fixed constant value range, it ensures that the control parameters are always within a safe and effective range, so that the traveling wheel can closely fit the conductor under different conductor spacing conditions, effectively reducing the risk of derailment. Attached Figure Description

[0019] Figure 1 A flowchart illustrating a method for optimizing the motion trajectory control of the working arm of an overhead maintenance robot, as provided in this application. Detailed Implementation

[0020] One embodiment of this application provides a method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot, as detailed in the following article. Figure 1 The method includes the following steps: S1: Determine the status parameters of the overhead maintenance robot during the current data collection cycle.

[0021] The overhead maintenance robot analyzed in this application includes at least four independent walking arms, which are evenly installed on both sides of the robot's body. Each walking arm is elastically hinged to a walking wheel, which is clamped onto the conductor of the overhead line. The opening of the walking arm is adjusted by the sliding mode control algorithm built into the overhead maintenance robot to control the movement of the walking wheel along the overhead conductor.

[0022] Specifically, the overhead maintenance robot analyzed in this embodiment includes four walking arms, each with four corresponding wheels. The four wheels of the overhead maintenance robot are respectively clamped onto the four conductors of the overhead four-split power line. During the robot's forward movement, two binocular depth cameras installed at the front and rear ends measure the distance between the overhead conductors where the front and rear wheels are located in real time, with a frame rate of 25 FPS. The two binocular depth cameras respectively capture images of the two front and two rear wheels of the overhead maintenance robot moving along the overhead conductors. The binocular depth cameras have a pre-trained target detection model built-in. The pre-trained target detection model can be an algorithm such as YOLOv8 or Faster R-CNN, without special restrictions; this embodiment uses YOLOv8.

[0023] Specifically, in this embodiment, the training process of the target detection model is as follows: 1000 historical images of the walking wheel moving along the overhead wire are acquired, and the contact area between the walking wheel and the wire (containing both the walking wheel and the wire) in the historical images is labeled using the LabelImg tool. The ground truth labels are divided into wheel-wire contact area and background. The labeled images are used as the training dataset for the YOLOv8 algorithm, where the optimizer is Adam, the loss function is cross-entropy loss function, the learning rate is 0.001, the batch-size is 8, and the epoch is set to 300. After training, the pre-trained target detection model is obtained.

[0024] Furthermore, the binocular camera uses images of the wheels running along the overhead conductor as input to a pre-trained target detection model to identify the wheel-line contact area between the wheels and the overhead conductor. The binocular depth camera uses built-in binocular stereo matching technology to obtain the 3D center coordinates of the wheel-line contact area and calculates the distance between the wheel-line contact areas of the front and rear wheels respectively, which is used as the measured distance between the front and rear wheels and the overhead conductor. The LabelImg tool, the trained target detection model, the binocular stereo matching algorithm of the binocular depth camera, and the target distance measurement of the binocular depth camera are well-known technologies, and their specific implementations will not be elaborated further.

[0025] Furthermore, a multi-dimensional force sensor (which can be a three-dimensional or six-dimensional force sensor, without special restrictions; this embodiment selects a six-dimensional force sensor) installed inside the wheels of the overhead maintenance robot is used to collect the forces acting on the wheels in three mutually perpendicular directions in real time. The x-axis points in the direction of the overhead maintenance robot's movement and is perpendicular to the y and z axes. The positive y-axis points perpendicular to the wheel surface and towards the left side of the overhead maintenance robot, while the positive z-axis points parallel to the vertical surface of the wheel and upwards. An angle sensor is installed at the hinge of the walking arm to collect the opening of the walking arm in real time. The sampling frequency is 25Hz, and the sampling period is 1s, meaning that N=25 data points are collected in one sampling period.

[0026] S2: Predict the distance between the overhead conductors where the front / rear traveling wheels are located in the next acquisition cycle; determine the conductor distance difference between the front / rear traveling wheels and the next acquisition cycle based on the predicted distance value. The conductor distance difference is used to measure the degree of fluctuation difference of the overhead conductors due to external factors between adjacent acquisition cycles.

[0027] When overhead maintenance robots move normally on overhead power lines, they are often affected by external factors such as strong winds, causing significant swaying of the overhead power line conductors. This alters the robot's trajectory, and failure to accurately and promptly adjust the trajectory can lead to derailment. Adjusting the robot's trajectory primarily uses a sliding mode control algorithm to adjust the pneumatic spring's extension and retraction to control the opening of the working arm (walking arm). However, traditional sliding mode control methods determine the approach rate based on fixed constants and sign functions. The fixed constants make it difficult for the approach rate to dynamically adapt to scenarios with significant external interference. For example, if the overhead power line sways violently due to strong winds, the sliding mode control will exhibit significant chattering, thus affecting the stability of the overhead maintenance robot's movement.

[0028] In order to analyze the degree of sway of the overhead conductor due to external factors in the next acquisition cycle, this application predicts the distance between the overhead conductors where the front and rear traveling wheels are located in the next acquisition cycle; based on the distance prediction value, the difference in conductor distance between the front and rear traveling wheels between the current acquisition cycle and the next acquisition cycle is determined, and the difference in conductor distance is used to measure the degree of fluctuation difference of the overhead conductor due to external factors between adjacent acquisition cycles.

[0029] Preferably, the present application predicts the distance between the overhead conductors where the front / rear wheels are located in the next acquisition cycle, including: obtaining the distances between the overhead conductors where the front / rear wheels are located in all historical acquisition cycles during the operation of the overhead maintenance robot, and forming a conductor distance sequence according to the acquisition time order; using the conductor distance sequence as the input of the ARIMA prediction algorithm to obtain the distance prediction values ​​between the overhead conductors where the front / rear wheels are located in the next acquisition cycle.

[0030] Preferably, the determination of the conductor distance difference in this application includes: calculating a first-order difference sequence of the sequence of predicted distances between the overhead conductors where the front / rear wheels are located in the next acquisition cycle, and obtaining the average value of the absolute values ​​of all elements in the first-order difference sequence; for the front / rear wheels of the overhead maintenance robot, determining the conductor distance difference between the front / rear wheels between the current acquisition cycle and the next acquisition cycle based on the difference between the average value calculated in the current acquisition cycle and the average value calculated in the next acquisition cycle.

[0031] Specifically, this embodiment takes the wheel-line contact area between the front wheels and the conductors on both sides of the overhead maintenance robot as an example. The distance between the overhead conductors of the front wheels collected in all historical collection cycles during the operation of the overhead maintenance robot is obtained. The distance sequence of the conductors is formed according to the collection time sequence. The distance sequence of the conductors is used as the input of the ARIMA prediction algorithm, wherein the autoregression order is 2, the difference order is 1, and the moving average order is 3. The sequence of predicted distances between the overhead conductors of the front wheels in the next collection cycle is obtained.

[0032] Next, the first-order difference sequence of the distance prediction values ​​for the next acquisition cycle is calculated. The more violent the conductor swings, the more the values ​​in the first-order difference sequence deviate from 0. Therefore, the average of the absolute values ​​of all elements in the first-order difference sequence is calculated. This reflects the average level of fluctuation in the distance between overhead conductors due to external factors, and is used in calculations. Using the same calculation principle, the average value of the first-order difference sequence of the distance prediction values ​​within the current acquisition period is calculated. . and The greater the difference, the greater the difference in the sway of the overhead conductor due to external factors in two adjacent acquisition cycles. This results in a greater difference in the distance fluctuation between the overhead conductors. Therefore, when performing sliding mode control in the next acquisition cycle, the fixed constant of the approach rate of the current acquisition cycle cannot be directly used. Instead, it should be adaptively adjusted according to the degree of interference from external factors in the two acquisition cycles to obtain a more reasonable approach rate for sliding mode control. The ARIMA prediction algorithm is a well-known technique, and its specific implementation will not be elaborated upon.

[0033] Furthermore, for the front wheel of the overhead maintenance robot, the difference in wire distance between the front wheel is determined based on the difference between the average value calculated in the current acquisition cycle and the average value calculated in the next acquisition cycle.

[0034] Specifically, in this embodiment, and The measure of the degree of difference is the difference in conductor distance. This indicates the difference in the guide wire distance between the front wheel and the next acquisition cycle. It is used to reflect the degree of fluctuation in the distance between overhead conductors due to external factors in two adjacent acquisition cycles. This is an error parameter to avoid the denominator being zero, which would render the calculation meaningless. Specifically, it can be determined according to... The value is determined by the dimension of the quantity, and in this embodiment, the value is 0.05. If the value is positive, it indicates that the overhead conductor will be significantly affected by external factors in the next acquisition cycle compared to the current acquisition cycle; conversely, if the value is negative, it indicates that the overhead conductor will be significantly affected by external factors in the next acquisition cycle. When the value is negative, it indicates that the overhead conductor will be less affected by external factors in the next acquisition cycle compared to the current acquisition cycle.

[0035] Use and Calculation Using the same calculation method, the second conductor distance sequence of the overhead conductors at the two rear walking wheels of the overhead maintenance robot is calculated, and then the conductor distance difference of the rear wheels is calculated. The specific process will not be elaborated further. The first and second conductor distance sequences are both denoted as conductor distance sequences.

[0036] S3: Determine the force interference degree of any walking wheel in the current acquisition cycle based on the vector force. The force interference degree is used to measure the degree of force fluctuation of the walking wheel due to external factors.

[0037] Further analysis reveals that when the distance between overhead power lines changes due to external factors, it typically alters the magnitude and direction of the force on the wheels of the overhead maintenance robot. For instance, an increase or decrease in the distance between overhead power lines usually results in a lateral pushing or pulling effect on the wheels, increasing the horizontal force. If the overhead maintenance robot cannot respond promptly or precisely adjust the opening of its arm, the distance between the wheels may not change with the distance between the overhead power lines. This typically increases friction between the wheels and the overhead power lines, hindering the movement of the wheels. Changes in the forces in three mutually perpendicular directions usually alter the magnitude and direction of the vector force on the wheels of the overhead maintenance robot.

[0038] To analyze the magnitude and direction of the vector force during the movement of each walking wheel of the overhead maintenance robot, taking the a-th walking wheel as an example, we obtain the forces in three mutually perpendicular directions collected at a certain sampling moment within the current acquisition period t. , , The vector force experienced by the a-th wheel at a certain sampling moment within the current acquisition period t. Equal to force The vector sum (resultant force), i.e. ,in, These represent the unit vectors along the x, y, and z axes, respectively. The vector sum of forces is a well-known technique, and its specific implementation will not be elaborated upon here.

[0039] When the overhead conductor is not disturbed by external factors and the distance between the conductors remains unchanged, the overhead maintenance robot moves steadily along the conductor. During the data collection period, the magnitude and direction of the force on each walking wheel are basically the same. Conversely, if the overhead conductor is seriously disturbed by external factors, the magnitude and direction of the vector force on the walking wheel of the overhead maintenance robot will usually fluctuate drastically.

[0040] In order to analyze the degree of interference of external factors on the force of the overhead maintenance robot's wheels, this application determines the force interference degree of any wheel in the current acquisition cycle based on the vector force. The force interference degree is used to measure the degree of force fluctuation of the wheel due to interference from external factors.

[0041] Preferably, the determination of the force interference degree in this application includes: obtaining the vector force experienced by any walking wheel at all sampling moments in the current acquisition period and forming a vector force sequence in chronological order; calculating the coefficient of variation of the magnitude of all vector forces in the vector force sequence; calculating the mean of the cosine similarity between adjacent vector forces in the vector force sequence; and determining the force interference degree of any walking wheel in the current acquisition period by performing a weighted summation based on the coefficient of variation and the mean of the cosine similarity.

[0042] Specifically, in this embodiment, the vector forces experienced by the a-th walking wheel at all sampling moments within the current acquisition period t are obtained and arranged into a vector force sequence according to time order. ,in, , , Let these represent the vector forces experienced by the a-th walking wheel at the 1st, 2nd, and Nth sampling times within the current acquisition period t; calculate the coefficient of variation of the magnitudes corresponding to all vector forces in the vector force sequence. This reflects the fluctuation of the force on the wheels of the overhead maintenance robot under external interference. Then, starting from the second vector force element in the vector force sequence, the cosine similarity *p* with the previous vector force element is calculated sequentially. Since the cosine similarity ranges from [-1, 1], a p value of -1 indicates that the two vector forces are in completely opposite directions, representing the lowest similarity, while a p value of 1 indicates that the two vector forces are in completely identical directions, representing the highest similarity. Therefore, to normalize *p* while maintaining its meaning, a method is used... This function normalizes all the cosine similarities and calculates the mean of all normalized cosine similarities. This reflects the similarity in the direction of force experienced by the wheels of the overhead maintenance robot under external interference. (Coefficient of Variation) The larger the mean of cosine similarity, the better. The smaller the value, the more significant the change in distance between overhead power lines under external interference during the data collection period, causing fluctuations in the magnitude and direction of the force on the wheels of the overhead maintenance robot. Therefore, it can be... and We perform a weighted summation to obtain the degree of force disturbance. , ;in, , These are the preset first weight and second weight, respectively. In this embodiment, the values ​​are 0.5. for Normalization function, This characterizes the degree to which the magnitude and direction of the force on the *a*-th traveling wheel of the overhead maintenance robot fluctuates due to external factors during the current data acquisition period. For error parameters, avoid Setting it to 0 renders subsequent calculations and analyses meaningless. The value range is 0.001-0.01, with no special restrictions. In this embodiment, the value is 0.001.

[0043] S4: Adjust the fixed constant of the sliding mode control approach rate for the next acquisition cycle based on the difference in conductor distance and the degree of force interference.

[0044] When an overhead maintenance robot uses sliding mode control to control the movement trajectory of its work arm in the next acquisition cycle, if the difference in the guide wire distance between the next and current acquisition cycles is significant, then the difference in the degree of force interference between the two cycles is also significant. In other words, the degree of interference from external factors differs considerably. Therefore, the fixed constant of the approach rate for sliding mode control in the next acquisition cycle cannot be the same as that of the current cycle, but should be adaptively adjusted according to the degree of force interference. Specifically, a negative guide wire distance difference indicates that the degree of interference from external factors in the next acquisition cycle is reduced, and the approach rate should be adjusted accordingly. The fixed constant is adjusted upwards, and the specific adjustment range is related to the degree of force interference. The greater the force interference, the smaller the upward adjustment range of the fixed constant; the smaller the force interference, the larger the upward adjustment range of the fixed constant. Conversely, if the difference in wire distance is positive, it indicates that the degree of interference from external factors in the next acquisition cycle is enhanced, and the fixed constant should be adjusted downwards. The greater the force interference, the larger the downward adjustment range of the fixed constant; the smaller the force interference, the smaller the downward adjustment range of the fixed constant. This is to avoid the fixed constant being too large, which would cause jitter in the sliding mode control, leading to frequent switching of the maintenance robot's motion trajectory and affecting the motion stability.

[0045] Based on the above analysis, this application adjusts the fixed constant of the sliding mode control approach rate for the next acquisition cycle according to the difference in conductor distance and the degree of force interference.

[0046] Preferably, this application obtains the current fixed constant of the sliding mode control algorithm approach rate for the current acquisition cycle; calculates the average value of the difference in guide distance between the front and rear wheels between the current acquisition cycle and the next acquisition cycle; determines the adjustment range based on the average value of the force interference of all walking wheels of the overhead maintenance robot in the current acquisition cycle; and determines the fixed constant of the adjusted sliding mode control approach rate for the next acquisition cycle based on the current fixed constant, the average value of the guide distance difference, and the adjustment range.

[0047] The method for determining the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle, based on the current fixed constant, the average value of the conductor distance difference, and the adjustment range, specifically includes: If the mean value of the conductor distance difference is greater than 0, the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle is reduced by the adjustment range based on the current fixed constant. If the mean value of the conductor distance difference is less than 0, the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle is increased by the adjustment range based on the current fixed constant. If the mean value of the conductor distance difference is equal to 0, then the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle remains unchanged as the current fixed constant.

[0048] Specifically, in this embodiment, the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle. The specific calculation method is as follows: In the formula, This is the current fixed constant of the approach rate of the sliding mode control algorithm in the current acquisition cycle. The unit of measurement is degrees per second. , This represents the average force disturbance experienced by the four wheels of the overhead maintenance robot during the current data collection cycle. This represents the maximum value normalization function, which is based on the maximum value among the average values ​​of the force disturbance of the four walking wheels during the historical data collection period of the overhead maintenance robot. Perform maximum value normalization when When it is greater than the maximum value, The value is assigned to 1; This represents the average difference in the guide wire distance between the front and rear wheels of the overhead maintenance robot between the current data acquisition cycle and its next acquisition cycle. ;when When the value is positive, it indicates that the overhead conductor is more significantly affected by external interference in the next acquisition cycle. In order to avoid chattering in sliding mode control, the fixed constant of the approach rate should be reduced. When the value is negative, it indicates that the overhead conductor is less affected by external interference in the next acquisition cycle. The fixed constant of the approach rate can be appropriately increased to increase the convergence speed of sliding mode control. When the value is 0, it indicates that the degree of interference from external factors on the overhead conductor is basically the same in the two consecutive acquisition cycles. Therefore, the fixed constant of the approach rate in the next acquisition cycle can continue to be the fixed constant of the current acquisition cycle.

[0049] It should be noted that the maximum value of the fixed parameter of the approach rate of the sliding mode control algorithm is 1, and the minimum value is 0.1. In this embodiment, the initial value is 0.5. If the fixed parameter after subsequent adaptive adjustment is greater than the maximum value or less than the minimum value, the fixed constant after adaptive adjustment is set to 1 or 0.1.

[0050] Based on the coefficient correction principle, the adjustment direction of the fixed constant is determined by the positive or negative relationship between the difference in wire distance between the next acquisition cycle and the current acquisition cycle. The average value of the force interference of the four wheels in the current acquisition cycle is used as the adjustment range. or As a correction coefficient, a fixed constant is used to adaptively adjust the sliding mode control approach rate in the next acquisition cycle.

[0051] S5: Based on the adjusted fixed constant, the opening of the traveling arm in the next acquisition cycle is controlled by the sliding mode control algorithm.

[0052] In the optimization control of the working arm motion trajectory of the overhead maintenance robot, the extension and contraction of the pneumatic spring is usually adjusted to control the opening of the robot's walking arm. This ensures that the walking wheels move closely along the overhead conductor and prevents the walking wheels from derailing due to changes in the distance between the overhead conductors caused by external factors.

[0053] Preferably, in this application, the sliding mode control algorithm controls the opening of the traveling arms in the next acquisition cycle, specifically including: determining the target distance between the front and rear traveling wheels in the next acquisition cycle based on the average predicted distance between the overhead conductors where the front and rear traveling wheels are located in the next acquisition cycle, and determining the target opening width of the front and rear traveling arms in the next acquisition cycle; determining the target opening of any traveling arm in the next acquisition cycle based on the target opening width; inputting the target opening of all traveling arms in the next acquisition cycle and the real-time opening of all traveling arms in the current acquisition cycle into the sliding mode control algorithm; the sliding mode control algorithm uses an adjusted fixed constant in the next acquisition cycle and outputs a control signal to adjust the extension and retraction of the pneumatic spring.

[0054] Specifically, the overhead maintenance robot of this application can adjust the distance between the left and right walking wheels from 400mm to 600mm. The minimum distance of 400mm between the left and right walking wheels is denoted as the frame width of the overhead maintenance robot. The target distance between the front and rear walking wheels is the average of the predicted distances between the overhead conductors where the front and rear walking wheels are located in the next acquisition cycle. Then, the target opening width of the left front and right front walking arms (or left rear and right rear walking arms) of the overhead maintenance robot in the next acquisition cycle is half of the difference between the target distance between the left front and right front walking wheels (or left rear and right rear walking wheels) and the frame width of the overhead maintenance robot in the next acquisition cycle.

[0055] Based on the law of cosines, the target opening width of each arm in the next acquisition cycle is obtained according to the target opening width of the arm. Taking the a-th arm as an example, the target opening width of the a-th arm is... And obtain the distance between the center of the walking wheel hinged to the a-th walking arm and the hinge point between the a-th walking arm and the robot. If is a known quantity that has been measured in advance, then the target opening of the a-th traveling arm is . ,in, It is the arcsine function. To find the minimum value function, As a preset constraint factor, to avoid A value greater than 1 will cause the calculation to exceed the limit. The value range is 0.95-1, and there are no special restrictions. In this embodiment, the value is 0.99.

[0056] Similarly, the target opening of the four arms of the overhead maintenance robot is obtained, in the order of left front, right front, left rear, and right rear. , , , The data, along with the real-time opening of the four traveling arms acquired during the current acquisition cycle, are used as inputs to the sliding mode control algorithm. The preset acquisition cycle for the sliding mode control algorithm is set to 0.04 s. The fixed constant of the sliding mode control approach rate within each acquisition cycle is an adaptively adjusted fixed constant for that acquisition cycle. That is, the sliding mode control algorithm uses the adjusted fixed constant in the next acquisition cycle. The approach rate is the constant velocity approach rate, and the sliding surface coefficient is 10. The PWM duty cycle output by the sliding mode control algorithm is transmitted to the working circuit controlling the pneumatic spring. The extension and contraction of the pneumatic spring are adjusted to make the opening of the traveling arm as close as possible to the target opening. This ensures that the traveling wheels move along the overhead conductor as closely as possible, resisting the swaying of the overhead conductor caused by external factors. Ultimately, this achieves optimized control of the working arm's motion trajectory of the overhead maintenance robot. The sliding mode control and the constant velocity approximation rate of the sliding surface are well-known techniques, and their specific implementation will not be elaborated upon.

Claims

1. A method for optimizing the motion trajectory of the working arm of an overhead maintenance robot, wherein the overhead maintenance robot comprises at least four independent walking arms, evenly installed on both sides of the robot's body, and the distance between the front / rear walking wheels, which are elastically hinged to the walking arms and located on both sides of the overhead maintenance robot, is adaptively controlled by a sliding mode control algorithm, characterized in that... The method includes the following steps: Determine the state parameters of the overhead maintenance robot during the current data acquisition cycle. The state parameters include the distance between overhead conductors, the vector force on the walking wheels, and the opening of the walking arm. Predict the distance between the overhead conductors where the front / rear traveling wheels are located in the next acquisition cycle; determine the conductor distance difference between the front / rear traveling wheels and the next acquisition cycle based on the distance prediction value. The conductor distance difference is used to measure the degree of fluctuation difference of the overhead conductors due to external factors between adjacent acquisition cycles. The force interference degree of any walking wheel in the current acquisition cycle is determined based on the vector force. The force interference degree is used to measure the degree of force fluctuation of the walking wheel due to external factors. The fixed constant of the sliding mode control approach rate for the next acquisition cycle is adjusted based on the difference in conductor distance and the degree of force interference. Based on the adjusted fixed constant, the opening of the traveling arm in the next acquisition cycle is controlled by the sliding mode control algorithm.

2. The method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot as described in claim 1, characterized in that, The determination of the distance between the overhead conductors includes: Images of the traveling wheel moving along the overhead conductor are captured using a binocular depth camera; The wheel-line contact area between the walking wheel and the overhead conductor in the image is identified using a pre-trained target detection model; The three-dimensional center coordinates of the wheel-line contact area are obtained using binocular stereo matching technology; The distance between the overhead conductors is calculated based on the three-dimensional center coordinates.

3. The method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot as described in claim 1, characterized in that, The predicted distance between the overhead conductors where the front / rear wheels are located in the next acquisition cycle includes: The distances between the front and rear walking wheels of the overhead conductors collected during all historical data collection periods during the operation of the overhead maintenance robot are obtained and arranged into a conductor distance sequence according to the collection time. The conductor distance sequence is used as input to the ARIMA prediction algorithm to obtain the predicted distance between the overhead conductors where the front and rear traveling wheels are located in the next acquisition cycle.

4. The method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot as described in claim 3, characterized in that, The determination of the difference in conductor distance includes: Calculate the first-order difference sequence of the sequence composed of the predicted distances between the overhead conductors where the front and rear wheels are located in the next acquisition cycle, and obtain the average value of the absolute values ​​of all elements in the first-order difference sequence; For the front and rear wheels of the overhead maintenance robot, the difference in the guide wire distance between the front and rear wheels between the current acquisition cycle and the next acquisition cycle is determined based on the difference between the average value calculated in the current acquisition cycle and the average value calculated in the next acquisition cycle.

5. The method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot as described in claim 1, characterized in that, The determination of the degree of force disturbance includes: Obtain the vector force experienced by any walking wheel at all sampling moments within the current acquisition period and assemble it into a vector force sequence in chronological order; Calculate the coefficient of variation of the magnitudes of all vector forces in the vector force sequence; Calculate the mean cosine similarity between adjacent vector forces in the vector force sequence; The force interference degree of any walking wheel in the current acquisition cycle is determined by weighted summation of the mean of the coefficient of variation and the cosine similarity.

6. The method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot as described in claim 4, characterized in that, The fixed constant for adjusting the sliding mode control approach rate of the next acquisition cycle based on the difference in conductor distance and the degree of force interference includes: Obtain the current fixed constant of the approach rate of the sliding mode control algorithm in the current acquisition period; Calculate the mean difference in guide wire distance between the front and rear wheels between the current acquisition cycle and the next acquisition cycle; The adjustment range is determined based on the average force interference of all the wheels of the overhead maintenance robot during the current data collection cycle. Based on the current fixed constant, the mean value of the conductor distance difference, and the adjustment range, determine the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle.

7. The method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot as described in claim 6, characterized in that, The step of determining the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle based on the current fixed constant, the mean value of the conductor distance difference, and the adjustment range includes: If the mean value of the conductor distance difference is greater than 0, the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle is reduced by the adjustment range based on the current fixed constant. If the mean value of the conductor distance difference is less than 0, the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle is increased by the adjustment range based on the current fixed constant. If the mean value of the conductor distance difference is equal to 0, then the fixed constant of the sliding mode control approach rate after adjustment in the next acquisition cycle remains unchanged as the current fixed constant.

8. The method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot as described in claim 1, characterized in that, The method of controlling the opening of the traveling arm in the next acquisition cycle through a sliding mode control algorithm includes: The target distance between the front and rear traveling wheels in the next acquisition cycle is determined based on the average value of the predicted distance between the overhead conductors where the front and rear traveling wheels are located in the next acquisition cycle, and the target opening width of the front and rear traveling arms in the next acquisition cycle is also determined. The target opening width is used to determine the target opening degree of any traveling arm in the next acquisition cycle; The target opening of all arms in the next acquisition cycle and the real-time opening of all arms in the current acquisition cycle are input into the sliding mode control algorithm. The sliding mode control algorithm uses an adjusted fixed constant in the next acquisition cycle and outputs a control signal to adjust the extension and retraction of the pneumatic spring.

9. The method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot as described in claim 6, characterized in that, The value range of the fixed constant after adjustment is 0.1 to 1. If the adjusted value exceeds the value range, the fixed constant is set as the boundary value of the value range.

10. The method for optimizing and controlling the motion trajectory of the working arm of an overhead maintenance robot as described in claim 8, characterized in that, The target opening width of the front / rear walking arms is half the difference between the target distance between the front / rear walking wheels and the width of the overhead maintenance robot frame.

Images