Underground high-precision robot control method and system based on machine vision
By employing a high-precision control method based on machine vision, utilizing sparse optical flow algorithm and nonlinear adaptive damping adjustment, the control stability problem of the downhole robotic arm under dust interference was solved, achieving rapid response and stable positioning, and improving the accuracy and safety of downhole operations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-04
- Publication Date
- 2026-04-10
AI Technical Summary
In the underground working environment, long cantilever robotic arms are difficult to control with high precision under dust interference. Traditional PID controllers have fixed parameters, which cannot balance rapid movement and stable hovering. They are also susceptible to visual signal interference, which can lead to malfunctions and poor system stability.
A high-precision control method based on machine vision is adopted. The feature points at the end of the robotic arm are tracked by the sparse optical flow algorithm, the consistency confidence of the visual flow field is calculated, the phase plane of position error and velocity error is constructed, and the damping coefficient is dynamically adjusted by the jitter energy divergence index and the nonlinear adaptive damping adjustment law to suppress the vibration of the robotic arm.
It enables rapid response and stable hovering of the robotic arm in complex downhole environments, significantly shortens the convergence time of residual vibration, and improves operational efficiency and safety.
Smart Images
Figure CN121821399A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of automated control technology for mining machinery. More specifically, this invention relates to a high-precision robot control method and system for underground mining based on machine vision. Background Technology
[0002] In modern mining operations, drilling rigs and shoveling rigs are core heavy equipment. Their actuators often employ robotic arms exceeding 5 meters in length, with long cantilever structures. These arms are primarily responsible for critical tasks such as clearing loose rocks from the roof and sidewalls of roadways after blasting operations and high-precision drilling. Due to the confined and narrow working spaces underground, these devices not only need to withstand the impact of extreme physical conditions but also face stringent requirements regarding the control precision, dynamic response speed, and continuous operation capability of the robotic arm's end effector.
[0003] Currently, the control of robotic arms in such devices mainly relies on traditional PID algorithms or manual remote control. Some high-end devices introduce machine vision, using cameras to assist in positioning. In actual operation, the robotic arm needs to move quickly to achieve efficient operation. Due to the significant flexibility of the long cantilever structure, when it stops abruptly at the target point or contacts the rock wall, the huge inertia and nonlinear pulsation of the hydraulic system will cause residual vibration at the end of the robotic arm. Traditional PID controllers have fixed parameters, making it difficult to balance the low damping required for rapid movement of the robotic arm with the high damping required for stable hovering.
[0004] However, residual smoke and dust from blasting, high concentrations of water mist, and splashing dust can easily interfere with the visual signals of optical sensors. This causes visual positioning algorithms to misidentify moving dust particles as feature points of the robotic arm. Such misidentification introduces a large amount of high-frequency random noise into the feedback displacement signal, leading to malfunctions or even systemic oscillations in the control system. This extremely harsh visual environment further exacerbates the difficulty of implementing automated control. In addition, existing control strategies mostly focus on correcting single position deviations, lacking comprehensive analysis and dynamic weight allocation of velocity change trends and perceived environmental confidence levels. This makes it impossible for the system to achieve adaptive intelligent degradation control when visual signals are impaired, and its overall stability is difficult to adapt to the complex and ever-changing downhole operational scenarios. Summary of the Invention
[0005] To address the technical challenges of downhole dust interfering with visual positioning signals and residual vibration in long cantilever robotic arms, this invention proposes a high-precision robot control method and system for downhole applications based on machine vision. This method can filter out environmental noise, dynamically adjust damping, and achieve high-precision control of the robotic arm, enabling it to resist vibration, respond quickly, and hover stably.
[0006] In a first aspect, the present invention provides a high-precision robot control method for downhole applications based on machine vision, comprising: acquiring an image of a region of interest at the end effector of a robotic arm; tracking the motion velocity vectors of all feature points within the region of interest; calculating statistical characteristics of the motion direction angles of all feature points; determining a visual flow field consistency confidence level based on the statistical characteristics, wherein the visual flow field consistency confidence level is negatively correlated with the dispersion of the motion direction angles of all feature points; using the visual flow field consistency confidence level to perform weighted fusion and filtering correction on the position error and velocity error of the end effector of the robotic arm, constructing a phase plane of position error and velocity error, and calculating a jitter energy divergence index based on the error parameters in the phase plane; calculating a real-time damping coefficient output to the controller based on the jitter energy divergence index using a nonlinear adaptive damping adjustment law, wherein the real-time damping coefficient increases with the increase of the jitter energy divergence index and continuously varies between a base damping value and a peak damping value; sending the real-time damping coefficient to the drive controller of the robotic arm, and suppressing the residual vibration of the robotic arm by adjusting the damping characteristics of the hydraulic or motor system, thereby achieving target positioning control of the robotic arm.
[0007] This invention effectively distinguishes between the rigid motion of the robotic arm and the disordered motion of dust by introducing statistical characteristic analysis of the physical optical flow field, thereby filtering out environmental noise interference in harsh environments. At the same time, by constructing a phase plane and combining it with dynamic adjustment of the real-time damping coefficient, the robotic arm can ensure rapid response with low damping when moving, and instantly switch to high damping to absorb mechanical energy when shaking or stopping. This method fundamentally solves the problem of adapting rapid response and stable positioning of long cantilever robotic arms in complex environments, significantly shortens the convergence time of residual vibration of the robotic arm, and improves work efficiency and safety.
[0008] Preferably, determining the confidence level of visual flow field consistency based on the statistical features includes: setting a marked area at the end of the robotic arm; acquiring images using an industrial camera at a preset frame rate, and cropping the acquired images to define the approximate range of the end of the robotic arm; simultaneously using a sparse optical flow algorithm to track the movement of all feature points within the region of interest, and obtaining the value of each feature point. exist The motion direction angle at any given time is determined based on the dispersion of the motion direction angles of all feature points within the region of interest. Confidence level of consistency between visual flow field at any time.
[0009] Preferably, the confidence level of visual flow field consistency The calculation method is as follows:
[0010] in, for The total number of all feature points tracked at any given time; for Time of the first The motion direction angle of each feature point, in degrees or radians; for The arithmetic mean of the motion direction angles of all feature points tracked at any given time; This is the sensitivity adjustment coefficient.
[0011] This invention is based on the physical law that the rigid body of the robotic arm moves in a highly consistent direction, while the dust moves in a chaotic manner. By evaluating the consistency of the movement direction of all feature points through standard deviation, it can accurately calculate the confidence level of the current visual flow field consistency. When the confidence level of the visual flow field consistency is high, visual information is fully utilized; when the confidence level of the visual flow field consistency is low, its weight is automatically reduced, thereby preventing misjudgment of the control system due to dust flying.
[0012] Preferably, the position and velocity errors of the robotic arm end effector are weighted, fused, and filtered using the visual flow field consistency confidence score, including: using the visual flow field consistency confidence score as a weight to perform Kalman filtering correction on the displacement and velocity of the robotic arm end effector, obtaining the position and velocity errors of the robotic arm end effector, constructing a phase plane of the position and velocity errors, and constructing a jitter energy divergence index based on the error parameters in the phase plane.
[0013] Preferably, the jitter energy divergence index satisfies the following relationship:
[0014] in, The vibration energy divergence index at the end of the robotic arm; for Confidence level of consistency between visual flow field at any time; for The distance error between the end effector position of the robotic arm and the target position at any given time; for The instantaneous speed of the robotic arm's end effector at any given moment; This is the weighting coefficient for the position error of the robotic arm's end effector; This is the weighting coefficient for the speed error at the end of the robotic arm; This is a nonlinear correction factor.
[0015] This invention no longer relies solely on the position error of the robotic arm's end effector, but instead integrates information from three dimensions—position, velocity, and environmental confidence—to construct a generalized function, particularly by introducing an exponential function. This makes the system highly sensitive to large positional errors at the end of the robotic arm, and can accurately identify the intensity of nonlinear jitter of the current robotic arm, providing a precise basis for subsequent strong intervention control.
[0016] Preferably, based on the jitter energy divergence index, the real-time damping coefficient output to the controller is calculated using a nonlinear adaptive damping adjustment law, including: setting a basic damping value and a peak damping value; when the jitter energy divergence index approaches zero, the real-time damping coefficient approaches the basic damping value; when the jitter energy divergence index increases, the real-time damping coefficient saturates and approaches the peak damping value.
[0017] Preferably, the real-time damping coefficient satisfies the following relationship:
[0018] in, This is the real-time damping coefficient; Based on the basic damping value, This represents the peak damping value. The adjustment sensitivity constant is greater than zero; It is an exponential function with the natural constant e as its base.
[0019] This invention employs a negative exponential saturation function model, ensuring that the real-time damping coefficient can transition continuously and smoothly between the base damping value and the peak damping value. This nonlinear adjustment method can quickly output damping close to the peak value to suppress vibration when high-energy jitter is detected, and can also avoid harmful physical impacts on the hydraulic system caused by sudden changes in damping, thus achieving a smooth and powerful control effect.
[0020] Preferably, constructing the phase plane of the position error and velocity error includes: constructing the phase plane with the position error as the horizontal axis and the velocity error as the vertical axis, so that the jitter trajectory of the end of the robotic arm is represented as a spiral around the origin on the phase plane.
[0021] Preferably, the calculation of the real-time damping coefficient output to the controller using the nonlinear adaptive damping adjustment law further includes: utilizing the characteristics of the negative exponential saturation function model to ensure that the real-time damping coefficient changes continuously and smoothly between the basic damping value and the peak damping value, so as to prevent a step change during the damping switching process from causing physical shock to the hydraulic system.
[0022] Secondly, the present invention provides a machine vision-based high-precision robot control system for downhole applications, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned machine vision-based high-precision robot control method for downhole applications is implemented.
[0023] By adopting the above technical solution, a computer program for a high-precision robot control method for downhole applications based on machine vision is generated and stored in a memory for loading and execution by a processor. This allows for the creation of a terminal device based on the memory and processor, facilitating its use.
[0024] The beneficial effects of this invention are as follows: This invention tracks feature points at the end of a robotic arm using a sparse optical flow algorithm. It calculates the consistency confidence of the visual flow field by utilizing the dispersion of the motion direction angle of the feature points, thus filtering out interference from downhole dust. Combining this visual flow field consistency confidence with position and velocity errors, a phase plane is constructed. The jitter energy divergence index is calculated to assess the nonlinear jitter intensity of the robotic arm. Then, a nonlinear adaptive adjustment law dynamically outputs the real-time damping coefficient to adjust the damping characteristics of the hydraulic or motor system, achieving high-precision control of the robotic arm for vibration resistance, fast response, and stable hovering.
[0025] Furthermore, by setting a highly reflective nano-bead reflective strip and a high frame rate industrial camera at the end of the robotic arm to improve image acquisition, and accurately calculating the visual flow field consistency confidence and jitter energy divergence index based on a statistical model, and using a negative exponential saturation function to ensure smooth and continuous adjustment of the real-time damping coefficient, the problems of fixed parameters in traditional PID control, susceptibility to dust interference, and difficulty in suppressing residual vibration of long cantilever structures in robotic arms are solved, significantly improving the accuracy, efficiency, and safety of robot operations under complex downhole conditions. Attached Figure Description
[0026] Figure 1 This is a flowchart illustrating a high-precision robot control method for downhole applications based on machine vision, as described in this invention. Figure 2 This is a comparison diagram of the vibration suppression response at the end of the robotic arm in an embodiment of the present invention; Figure 3 This is a plane trajectory comparison diagram of the position error and velocity error of the robotic arm end in an embodiment of the present invention. Detailed Implementation
[0027] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.
[0028] This invention discloses a machine vision-based high-precision robot control method for downhole applications, referring to... Figure 1 This includes steps S1-S4: S1. Obtain an image of the region of interest at the end of the robotic arm, track the motion velocity vectors of all feature points within the region of interest, calculate the statistical characteristics of the motion direction angles of all feature points, and determine the visual flow field consistency confidence level based on the statistical characteristics. The visual flow field consistency confidence level is negatively correlated with the dispersion of the motion direction angles of all feature points.
[0029] In an optional embodiment, a marking area is set at the end of the robotic arm to cope with the low-light environment downhole. The marking area can be a highly reflective nanobead reflective strip, using a recommended frame rate of not less than [missing information]. Industrial cameras acquire images, and regions of interest (ROIs) are cropped within these images to approximate the extent of the robotic arm's end effector, thereby reducing computational overhead caused by irrelevant backgrounds. Simultaneously, sparse optical flow algorithms, such as the Lucas-Kanade method, are used to track the motion of all feature points within the ROI. Tracked in real time Each feature point is obtained. Calculate the velocity vector of the object and its direction angle. .
[0030] Since the robotic arm is a rigid body, the motion directions of all feature points on its surface should be highly consistent. However, dust, as an aerosol, is affected by airflow, and its motion directions are chaotic. Therefore, the degree of dispersion of the motion direction angles of all feature points in the region of interest at the end of the robotic arm can be calculated to determine the direction of motion. Confidence of visual flow field consistency at any time The confidence level of visual flow field consistency is calculated as follows:
[0031] in, for The total number of all feature points tracked at any given time; for Time of the first The motion direction angle of each feature point, in degrees or radians; for The arithmetic mean of the motion direction angles of all feature points tracked at any given time; This is the sensitivity adjustment coefficient, for example, a value of 0.5.
[0032] For example, suppose The system tracks N=4 feature points at all times, with a sensitivity adjustment coefficient λ=0.5.
[0033] Assuming a clean and dust-free underground environment, all feature points are points on the robotic arm, moving in the same direction and height. Let's assume the angles of the four feature points are... ,but The arithmetic mean of the motion direction angles of all feature points tracked at any given time Variance calculation term Standard deviation Finally, the confidence level of visual flow field consistency was obtained. The result is close to 1, indicating that the visual signal is highly reliable.
[0034] In the presence of dust interference in the underground environment, the characteristic points are the points on the robotic arm, upward-flying dust, and downward-drifting dust. Assume the angles of the four characteristic points are as follows: ,but The arithmetic mean of the motion direction angles of all feature points tracked at any given time ,because and If the deviation from the mean is large, the variance term will be very large. Assuming the standard deviation is 40, the final confidence level of visual flow field consistency will be obtained. The result is close to 0, indicating that the visual signal is unreliable.
[0035] Thus, by analyzing the motion direction angles of all feature points and calculating the confidence level of visual flow field consistency, the true motion state of the robotic arm end can be accurately obtained in the dusty underground environment, eliminating false signals caused by dust and ensuring the purity of the input source of the control system.
[0036] S2. The position and velocity errors at the end of the robotic arm are weighted, fused, and filtered using the visual flow field consistency confidence level. A phase plane of position and velocity errors is constructed, and the jitter energy divergence index is calculated based on the error parameters in the phase plane.
[0037] In an optional embodiment, the visual flow field consistency confidence score calculated by S1 is used. As weights, Kalman filtering is applied to correct the displacement and velocity of the robotic arm's end effector. When the confidence level of the visual flow field consistency is low, the filter relies more on the system's predictive model than on visual measurements, thus obtaining an accurate end-effector position error. and speed error and with The horizontal axis is... Construct a phase plane for the vertical axis.
[0038] Specifically, the jitter trajectory of the robotic arm's end effector appears as a spiral around the origin in the phase plane. By analyzing the dynamic changes of the error trajectory within the phase plane, the motion trend of the robotic arm can be effectively identified, ensuring that the system can obtain a specific numerical standard. This provides a precise data foundation for subsequent control decisions, thereby constructing a jitter energy divergence index, which satisfies the following relationship:
[0039] in, The vibration energy divergence index at the end of the robotic arm; for Confidence level of consistency between visual flow field at any time; for The distance error between the end effector position of the robotic arm and the target position at any given time; for The instantaneous speed of the robotic arm's end effector at any given moment; This is the weighting coefficient for the position error of the robotic arm's end effector; This is the weighting coefficient for the speed error at the end of the robotic arm; This is a nonlinear correction factor; for example, it can take the following values: .
[0040] For example, suppose the end-effector position error of the robotic arm is corrected by filtering. meters, speed error meters per second, weighting factor for position error Weighting coefficient of speed error Nonlinear correction factor .
[0041] Given a clean and dust-free downhole environment, the current confidence level of visual flow field consistency is... This leads to the divergence index of the vibration energy at the end of the robotic arm. At this point, the index reflects the actual physical vibration at the end of the robotic arm.
[0042] Given the presence of dust interference in the downhole environment, the current confidence level of visual flow field consistency is... This yields the vibration energy divergence index at the end of the robotic arm. At this time, the energy index will be suppressed, which can effectively prevent the system from mistakenly thinking that the robotic arm is shaking violently due to dust, thus avoiding erroneous forced control actions.
[0043] Among them, nonlinear correction factor The design of this feature results in a relatively large positional error at the end of the robotic arm. For example, when the error reaches 0.5 meters, the first term will increase exponentially, causing the vibration energy at the end of the robotic arm to diverge exponentially. The rapid increase reflects the control logic that large deviations require strong intervention.
[0044] Thus, by constructing the phase plane trajectory and jitter energy divergence index of the robotic arm's end effector, the system no longer relies on the end position deviation of the robotic arm for judgment, but combines speed information and confidence level for comprehensive evaluation, thereby achieving effective suppression of the nonlinear jitter intensity of the robotic arm's end effector.
[0045] S3. Based on the jitter energy divergence index, the real-time damping coefficient output to the controller is calculated using a nonlinear adaptive damping adjustment law. This real-time damping coefficient increases with the increase of the jitter energy divergence index and changes continuously between the basic damping value and the peak damping value.
[0046] In an optional embodiment, based on the jitter energy divergence index calculated by S2, when the jitter energy divergence index is small, i.e., the system tends to a steady state or the robotic arm is in a normal moving state, the system outputs a small real-time damping coefficient to ensure the robotic arm's movement is sensitive; when the jitter energy divergence index is large, i.e., the robotic arm jitters or experiences residual vibration due to a sudden stop, the system outputs a very large real-time damping coefficient, thereby forcibly absorbing mechanical energy and ensuring the robotic arm quickly reaches a stable state; the real-time damping coefficient satisfies the following relationship:
[0047] in, This is the real-time damping coefficient; Based on the basic damping value, This represents the peak damping value. The adjustment sensitivity constant is greater than zero; It is an exponential function with the natural constant e as its base.
[0048] For example, assume basic damping Ns / m, peak damping Ns / m, adjustment constant .
[0049] When the system tends to a steady state or the robotic arm is in a normal moving state, the jitter energy divergence index is... At this time, the real-time damping coefficient The robotic arm experiences minimal damping, resulting in smooth movement.
[0050] When a robotic arm vibrates or stops abruptly, resulting in residual vibration, the vibration energy divergence index is... At this point, the exponential function Then the real-time damping coefficient The damping experienced by the robotic arm is close to the peak damping, which quickly suppresses the vibration of the robotic arm and brings it to a stable state.
[0051] When the robotic arm is in a transitional state, the jitter energy divergence index At this point, the exponential function Then the real-time damping coefficient The coefficient is calculated using a negative exponential saturation function model to ensure that the real-time damping coefficient changes smoothly and continuously between the base damping value and the peak damping value.
[0052] In this way, by using a nonlinear adaptive damping adjustment law, a control effect that combines both soft and hard approaches is achieved, ensuring smooth operation while also possessing strong anti-disturbance and vibration suppression capabilities.
[0053] S4. The real-time damping coefficient is sent to the drive controller of the robotic arm. The residual vibration of the robotic arm is suppressed by adjusting the damping characteristics of the hydraulic or motor system, thereby achieving the target positioning control of the robotic arm.
[0054] In an optional embodiment, the execution control module converts the calculated real-time damping coefficient into control commands in real time and sends them to the hydraulic servo valve controller or motor driver to dynamically change the damping characteristics of the actuator.
[0055] like Figure 2 The figure shown is a comparison of the vibration suppression response of the robotic arm end effector in an embodiment of the present invention. It can be seen that the traditional fixed PID control in the prior art exhibits a large-amplitude sinusoidal decay waveform and a long vibration duration, indicating that the robotic arm repeatedly swayed after reaching its position. In contrast, the adaptive damping control of the present invention shows that there is almost no reverse overshoot after the initial disturbance, but rather a rapid and smooth return to zero. This indicates that the present invention successfully triggers a strong damping mechanism when a large deviation is detected, which can dissipate the vibration energy instantly.
[0056] like Figure 3 The figure shows a comparison of the phase plane trajectories of the position error and velocity error at the end of the robotic arm in an embodiment of the present invention. It can be seen that the trajectory of the prior art presents a wide and dense spiral shape, indicating that the robotic arm vibrates for a long time in both the position and velocity dimensions. In contrast, the convergence trajectory of the present invention presents a very compact curve that quickly wraps around the origin, and the arrow of the steady-state target point in the figure points to the origin of the coordinate system, indicating that the present invention can drive the system to return to a stable state with the shortest path and the least energy consumption.
[0057] In this way, by adjusting the control module in real time, high-precision and high-stability control of the downhole heavy robotic arm was finally achieved.
[0058] This invention also discloses a machine vision-based high-precision robot control system for downhole applications, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a machine vision-based high-precision robot control method for downhole applications according to this invention.
[0059] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.
[0060] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.
Claims
1. A high-precision robot control method for downhole applications based on machine vision, characterized in that, include: The image of the region of interest at the end of the robotic arm is acquired, the motion velocity vectors of all feature points within the region of interest are tracked, the statistical characteristics of the motion direction angles of all feature points are calculated, and the visual flow field consistency confidence is determined based on the statistical characteristics. The visual flow field consistency confidence is negatively correlated with the dispersion of the motion direction angles of all feature points. The position and velocity errors at the end of the robotic arm are weighted, fused, and filtered using the visual flow field consistency confidence level. A phase plane of position and velocity errors is constructed, and the jitter energy divergence index is calculated based on the error parameters in the phase plane. Based on the jitter energy divergence index, the real-time damping coefficient output to the controller is calculated using a nonlinear adaptive damping adjustment law. The real-time damping coefficient increases with the increase of the jitter energy divergence index and changes continuously between the base damping value and the peak damping value. The real-time damping coefficient is sent to the drive controller of the robotic arm, and the residual vibration of the robotic arm is suppressed by adjusting the damping characteristics of the hydraulic or motor system, thereby achieving target positioning control of the robotic arm.
2. The high-precision robot control method for downhole applications based on machine vision according to claim 1, characterized in that, Determining the confidence level of visual flow field consistency based on the statistical characteristics includes: setting a marked area at the end of the robotic arm; acquiring images using an industrial camera at a preset frame rate, and cropping the acquired images to define the approximate range of the robotic arm's end; and simultaneously using a sparse optical flow algorithm to track the movement of all feature points within the region of interest, obtaining the value of each feature point. exist The motion direction angle at any given time is determined based on the degree of dispersion of the motion direction angles of all feature points within the region of interest. Confidence level of consistency between visual flow field at any time.
3. The high-precision robot control method for downhole applications based on machine vision according to claim 2, characterized in that, Confidence of visual flow field consistency at any time The calculation method is as follows: in, for The total number of all feature points tracked at any given time; for Time of the first The motion direction angle of each feature point, in degrees or radians; for The arithmetic mean of the motion direction angles of all feature points tracked at any given time; This is the sensitivity adjustment coefficient.
4. The high-precision robot control method for downhole applications based on machine vision according to claim 1, characterized in that, The position and velocity errors of the robotic arm end effector are weighted, fused, and filtered using the visual flow field consistency confidence score. This includes: using the visual flow field consistency confidence score as a weight to perform Kalman filtering correction on the displacement and velocity of the robotic arm end effector, obtaining the position and velocity errors of the robotic arm end effector, constructing a phase plane of the position and velocity errors, and constructing a jitter energy divergence index based on the error parameters in the phase plane.
5. The high-precision robot control method for downhole applications based on machine vision according to claim 4, characterized in that, The jitter energy divergence index satisfies the following relationship: in, The vibration energy divergence index at the end of the robotic arm; for Confidence level of consistency between visual flow field at any time; for The distance error between the end effector position of the robotic arm and the target position at any given time; for The instantaneous speed of the robotic arm's end effector at any given moment; This is the weighting coefficient for the position error of the robotic arm's end effector; This is the weighting coefficient for the speed error at the end of the robotic arm; This is a nonlinear correction factor.
6. The high-precision robot control method for downhole applications based on machine vision according to claim 5, characterized in that, Based on the jitter energy divergence index, the real-time damping coefficient output to the controller is calculated using a nonlinear adaptive damping adjustment law, including: setting a basic damping value and a peak damping value; when the jitter energy divergence index approaches zero, the real-time damping coefficient approaches the basic damping value; when the jitter energy divergence index increases, the real-time damping coefficient saturates and approaches the peak damping value.
7. A high-precision robot control method for downhole applications based on machine vision according to claim 6, characterized in that, The real-time damping coefficient satisfies the following relationship: in, This is the real-time damping coefficient; Based on the basic damping value, This represents the peak damping value. The adjustment sensitivity constant is greater than zero; It is an exponential function with the natural constant e as its base.
8. A high-precision robot control method for downhole applications based on machine vision according to claim 4, characterized in that, Constructing the phase plane of the position error and velocity error includes: constructing the phase plane with the position error as the horizontal axis and the velocity error as the vertical axis, so that the jitter trajectory of the end of the robotic arm is represented as a spiral around the origin on the phase plane.
9. A high-precision robot control method for downhole applications based on machine vision according to claim 7, characterized in that, The method of calculating the real-time damping coefficient output to the controller using a nonlinear adaptive damping adjustment law also includes: using the characteristics of a negative exponential saturation function model to ensure that the real-time damping coefficient changes continuously and smoothly between the basic damping value and the peak damping value, so as to prevent step abrupt changes during damping switching that could cause physical shocks to the hydraulic system.
10. A high-precision robot control system for downhole applications based on machine vision, characterized in that, include: The system includes a processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement a high-precision robot control method for downhole applications based on machine vision, as described in any one of claims 1-9.