Insulation mechanical arm self-adaptive control method for hot-line work platform

Abstract

The invention relates to the technical field of industrial control systems, and discloses an insulating mechanical arm self-adaptive control method for a hot-line work platform. The method comprises the following steps: acquiring joint angle, angular velocity, torque, temperature, humidity and wind speed information in real time; time-varying stiffness and damping parameters are identified on line based on a recursive least square method with a forgetting factor; constructing a nonlinear disturbance observer to estimate a composite disturbance torque and generate feed-forward compensation; dynamically adjusting the PID gain according to a pole assignment strategy; and fusing the feedback control torque, the model feed-forward torque and the disturbance compensation torque to generate a synthetic control instruction. According to the invention, through a dual adaptive mechanism, the trajectory tracking precision, interference resistance and dynamic stability of the mechanical arm under complex working conditions are improved.

Description

Technical Field

[0001] This invention belongs to the field of industrial control system technology, specifically relating to an adaptive control method for an insulated robotic arm used in a live-line working platform. Background Technology

[0002] With the continuous improvement of automation in power system operation and maintenance, live-line working platforms play a crucial role in high-voltage transmission line maintenance, equipment installation, and fault handling. As the core actuator of a live-line working platform, the insulated robotic arm must achieve high-precision and high-safety remote operation in a strong electric field environment. To ensure the safety of personnel and equipment, its structure generally uses composite insulating materials to manufacture joint components to meet electrical isolation requirements.

[0003] However, during long-term service, the mechanical properties (such as equivalent stiffness and damping) of such materials are affected by factors such as changes in environmental temperature and humidity, mechanical fatigue and aging, which will cause a slow but not negligible time-varying drift, resulting in a mismatch in the system dynamics model.

[0004] Motion control of insulated robotic arms is highly dependent on accurate modeling of their joint dynamic characteristics. Traditional industrial control strategies often employ fixed-parameter PID controllers, achieving trajectory tracking through preset gains. This method performs well under ideal conditions where system parameters are constant and external disturbances are negligible. However, in actual live-line working scenarios, robotic arms not only face the continuous evolution of their own material properties but also need to cope with complex external disturbances such as sudden changes in wind load and end-effector load switching.

[0005] The combined uncertainty of time-varying internal parameters and external disturbances makes it difficult for fixed-gain controllers to maintain stable performance. This often manifests as increased trajectory tracking error, aggravated joint jitter, or even control instability, which severely restricts operational accuracy and safety.

[0006] While some existing technologies attempt to introduce feedforward compensation or robust control to suppress disturbances, their adaptability to parameter drift is limited, and they lack online sensing mechanisms for joint equivalent mechanics parameters. Furthermore, most methods fail to establish a reference benchmark for ideal dynamic response, resulting in a lack of clear objectives for controller adjustments and making it difficult to achieve smooth and precise motion output while ensuring stability.

[0007] Especially in harsh working environments such as high altitude, strong winds, and high humidity, the above-mentioned defects are further amplified, resulting in a decrease in the reliability and working efficiency of the insulated robotic arm.

[0008] Therefore, there is an urgent need for a new control method that can identify joint dynamic parameters in real time and adaptively adjust the control law based on a reference model, in order to jointly address the dual challenges of internal degradation and external disturbances. Summary of the Invention

[0009] The technical problem to be solved by this invention is that the existing control method for insulated robotic arms used in live-line working platforms adopts a proportional-integral-derivative controller with fixed parameters. This controller cannot effectively cope with the slow time-varying characteristics caused by changes in the stiffness and damping parameters of composite material joints over time and with changes in ambient temperature and humidity. It also cannot suppress the strong external disturbances introduced by sudden changes in wind load and end-effector load during operation. This results in a decrease in the trajectory tracking accuracy of the robotic arm and the generation of high-frequency jitter, which affects the safety and efficiency of the operation.

[0010] To address the aforementioned technical problems, this invention provides an adaptive control method for an insulated robotic arm used in live-line working platforms. This method constructs a dual-decoupled control framework to identify the core dynamic parameters of the robotic arm joints in real time and online. Based on this identification result, a composite disturbance observer is simultaneously constructed to accurately estimate and feedforward compensate for external disturbances. At the same time, the gain parameters of the main controller are dynamically adjusted, thereby achieving dual adaptation to the time-varying internal parameters of the robotic arm and external environmental disturbances. This ensures that the robotic arm maintains high-precision trajectory tracking performance and dynamic stability throughout its entire life cycle and under complex working conditions.

[0011] According to an aspect of the present invention, an adaptive control method for an insulated robotic arm for a live-line working platform is provided, comprising the following steps: The status information of each joint of the insulated robotic arm is acquired in real time. The joint status information includes joint angle, joint angular velocity, joint driving torque, joint body temperature, ambient humidity around the joint, and wind speed vector at the location of the live working platform. A second-order lumped parameter dynamic model of a single joint of the insulated robotic arm is established, and a recursive least squares algorithm with a forgetting factor is used to identify and continuously update the time-varying stiffness parameters and time-varying damping parameters in the dynamic model online based on the joint state information acquired in real time. Based on the time-varying stiffness parameters and time-varying damping parameters identified online, a nonlinear disturbance observer is constructed for estimating the composite disturbance moment, which includes the equivalent disturbance moment caused by wind load, end load changes, and dynamics not modeled in the model. Using the nonlinear disturbance observer, the estimated value of the composite disturbance torque acting on each joint is calculated in real time, and a feedforward compensation torque that is equal in magnitude and opposite in direction to the estimated value is generated. Based on the time-varying stiffness and damping parameters identified online, and according to the preset closed-loop system pole configuration strategy, the proportional gain, integral gain, and derivative gain of the proportional-integral-derivative main controller are dynamically and in real time calculated and adjusted. Based on the desired joint motion trajectory, and combined with the time-varying stiffness and damping parameters identified online, the theoretical feedforward torque used to counteract the joint's own dynamics is calculated. The feedback control torque calculated by the proportional-integral-derivative main controller based on the trajectory tracking error, the theoretical feedforward torque, and the feedforward compensation torque are algebraically summed to generate the final synthetic control torque applied to the joint drive unit, which drives the insulated robotic arm joint to accurately track the desired joint motion trajectory.

[0012] As one embodiment of the present invention, the real-time acquisition of the state information of each joint of the insulated robotic arm specifically includes: The real-time angle of the joint is obtained by using a multi-turn absolute encoder with an accuracy better than 0.01 arcsecond installed at the output end of each joint. The real-time angular velocity of the joint is obtained by performing a first-order difference operation on the joint angle signal and filtering it through a low-pass digital filter with a cutoff frequency of 100 Hz. The real-time driving torque applied to the joint is obtained by deploying a strain gauge torque sensor between the joint drive motor and the reducer. The real-time temperature of the joint body is obtained by a platinum resistance temperature sensor that is closely attached to the surface of the joint composite material shell. The ambient humidity around the joint is obtained by using a capacitive humidity sensor installed on the base of the robotic arm; A three-dimensional ultrasonic anemometer installed on the top of the live-line working platform is used to obtain the three-dimensional wind speed vector at the platform's location. Based on the real-time configuration of the robotic arm, the wind speed vector is calculated into the coordinate system of each joint.

[0013] As one embodiment of the present invention, the step of establishing a second-order lumped parameter dynamic model of a single joint of an insulated robotic arm and performing online identification using a recursive least squares algorithm with a forgetting factor specifically includes: The dynamic behavior of a single joint can be described by the following differential equation: ; in Let be the moment of inertia of the joint. , and These are the joint angle, angular velocity, and angular acceleration, respectively. and For the time-varying damping parameters and time-varying stiffness parameters to be identified, For driving torque, It is a composite disturbance moment; Transform the above differential equation into a discrete-time linear regression form: ; Sampling period; regression vector ; Parameter vector to be identified ; The parameter vector Θ(k) is updated recursively using a recursive least squares algorithm with a forgetting factor, and its update law is: ; ; ; in for The estimated values ​​of the identification parameters at time points. Here is the gain matrix. Let covariance matrix be the variance matrix. The forgetting factor is set to a value range of 0.98 to 0.995.

[0014] As one embodiment of the present invention, the construction of the nonlinear disturbance observer for estimating the composite disturbance moment specifically includes: Define state variables Its dynamic equation is: ; in For observer gain, and The estimated values ​​of time-varying damping and stiffness parameters output by the online identification module; The estimated value of the composite disturbance moment It is calculated using the following formula: ; The observer gain Designed as a positive definite diagonal matrix, the diagonal element values ​​are related to the time-varying stiffness parameter. It is proportional to the square root to ensure that the observer bandwidth can adaptively adjust as the system stiffness changes.

[0015] As one embodiment of the present invention, the step of dynamically and in real-time calculating and adjusting the gain parameters of the proportional-integral-derivative main controller according to a preset closed-loop system pole configuration strategy specifically includes: The desired closed-loop control system has a characteristic equation in second-order standard form: ; in For the desired damping ratio, The desired undamped natural frequency, the desired damping ratio The desired undamped natural frequency is set to 0.707. Select within the range of 10 to 30 radians per second, depending on the requirements of the task. The proportional gain of the proportional-integral-derivative main controller Integral gain With differential gain Real-time calculations are performed based on the following formula: ; ; ; in The integral gain adjustment coefficient has a value ranging from 0.1 to 0.3.

[0016] As one embodiment of the present invention, the steps of calculating the theoretical feedforward torque and generating the final composite control torque specifically include: Obtain the desired joint motion trajectory and its first derivative and second derivative ; Calculate the theoretical feedforward torque: ; Calculate the trajectory tracking error: ; Calculate the feedback control torque of the proportional-integral-derivative main controller: ; Generate feedforward compensation torque: ; Calculate the final resultant control torque: ; The torque command is then sent to the joint's servo drive.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. By introducing a recursive least squares algorithm with a forgetting factor, the precise online identification of the time-varying stiffness and damping parameters of the composite material joints of the insulated robotic arm is realized. This enables the control system to perceive and adapt to the drift of internal physical properties caused by material aging, temperature and humidity changes in real time, fundamentally solving the performance degradation problem caused by model mismatch in traditional fixed parameter controllers.

[0018] 2. A nonlinear disturbance observer based on real-time updated model parameters was constructed, which can accurately estimate and actively compensate for the complex disturbances caused by wind load, changes in end-effector load, and unmodeled dynamics. This greatly enhances the robot arm's resistance to disturbances and robustness in harsh external environments and suppresses vibration.

[0019] 3. An adaptive gain adjustment mechanism based on pole placement strategy is adopted to ensure that the closed-loop dynamic response characteristics of the robotic arm are always maintained at the preset, optimal operating point, regardless of changes in the internal parameters of the robotic arm, thus guaranteeing the consistency of the speed, accuracy and stability of trajectory tracking.

[0020] 4. By organically combining adaptive feedback control, model feedforward control, and disturbance feedforward compensation, a composite control law with multiple compensations is formed. The synergistic effect enables the system to not only quickly correct errors that have occurred, but also to predictively offset most of the predictable dynamic forces and external disturbances, thereby reducing the generation of tracking errors at the source and achieving unprecedented high-precision motion control. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall technical solution architecture of the adaptive control method for an insulated robotic arm for a live-line working platform proposed in this invention. Figure 2 This is a schematic diagram of the core principle framework of the dual decoupling adaptive control framework in this invention; Figure 3 This is a flowchart illustrating the logical process of online identification of joint time-varying parameters and dynamic model update in this invention. Figure 4 This is a flowchart illustrating the logical flow of the construction and feedforward compensation of the nonlinear disturbance observer based on real-time model parameters in this invention. Figure 5 This is a flowchart illustrating the logical flow of the adaptive gain adjustment of the PID controller based on the pole placement strategy in this invention. Figure 6 This is a schematic diagram of the multi-level interaction relationship and data flow between the terminal joint driving unit and the multi-source sensor information fusion in this invention. Detailed Implementation

[0022] Please refer to Figures 1 to 6 This invention provides an adaptive control method for an insulated robotic arm used in live-line working platforms, aiming to solve the problems of decreased trajectory tracking accuracy and high-frequency jitter caused by the slow drift of joint stiffness and damping parameters of composite materials with time, temperature and humidity, as well as the sudden changes in wind load and end load during operation.

[0023] This method constructs a dual-decoupled adaptive control framework to achieve online identification of the time-varying characteristics of the internal dynamic parameters of the robotic arm and accurate estimation of external complex disturbances. Based on this, the gain of the main controller is dynamically adjusted to generate a synthetic control torque that includes a triple compensation mechanism of feedback, model feedforward and disturbance feedforward. This ensures that the insulated robotic arm has high-precision trajectory tracking capability and dynamic stability throughout its entire life cycle and under complex working conditions.

[0024] The method includes the following steps: acquiring the status information of each joint of the insulated robotic arm in real time; A second-order lumped parameter dynamic model of a single joint is established, and a recursive least squares algorithm with a forgetting factor is used to identify time-varying stiffness parameters and time-varying damping parameters online. A nonlinear disturbance observer is constructed based on the identification results to estimate the composite disturbance moment; Generate a feedforward compensation torque that is equal in magnitude and opposite in direction to the disturbance estimate; dynamically calculate the proportional gain, integral gain and derivative gain of the proportional-integral-derivative main controller based on the identified time-varying parameters and the preset closed-loop pole configuration strategy; The theoretical feedforward torque is calculated by combining the expected trajectory; The feedback control torque, theoretical feedforward torque, and feedforward compensation torque are algebraically summed to generate the final synthetic control torque applied to the joint drive unit.

[0025] First, step S1 is executed to obtain the status information of each joint of the insulated robotic arm in real time.

[0026] This status information includes joint angle, joint angular velocity, joint driving torque, joint body temperature, ambient humidity around the joint, and wind speed vector at the location of the live-line working platform.

[0027] The joint angles are obtained by a multi-turn absolute encoder installed at the output end of each joint. The encoder has a measurement accuracy better than 0.01 arcseconds and can provide absolute position information without cumulative error.

[0028] The joint angular velocity is obtained by performing a first-order differential operation on the joint angle signal, and then filtered by a low-pass digital filter with a cutoff frequency of 100 Hz to suppress the amplification effect of high-frequency noise on the differential operation and ensure the smoothness and accuracy of the angular velocity signal.

[0029] The joint drive torque is directly measured by a strain gauge torque sensor deployed between the joint drive motor and the reducer. This sensor has high linearity and temperature stability, and its sampling frequency is greater than 1000 Hz to match the time response requirements of the control system.

[0030] The joint body temperature is obtained by a platinum resistance temperature sensor that is tightly attached to the surface of the joint composite material shell. The sensor is encapsulated in a thermally conductive silicone grease layer to ensure good thermal contact with the shell, and the temperature measurement response time is less than 500 milliseconds.

[0031] The ambient humidity is collected by a capacitive humidity sensor installed on the base of the robotic arm. This sensor has an anti-condensation structure and can work stably for a long time in high humidity environments.

[0032] The three-dimensional wind speed vector is measured in real time by a three-dimensional ultrasonic anemometer installed on the top of the live-line working platform. This anemometer has no moving parts, strong anti-electromagnetic interference capability, and is suitable for high-voltage live-line environments.

[0033] The obtained global wind speed vector needs to be solved step by step into the local coordinate system of each joint through a homogeneous transformation matrix based on the current configuration of the robotic arm—that is, the pose vector composed of the angles of each joint—so as to obtain the effective wind load direction and magnitude acting on each joint, providing input basis for subsequent disturbance modeling.

[0034] Then, step S2 is executed to establish a second-order lumped parameter dynamic model of a single joint of the insulated robotic arm, and a recursive least squares algorithm with a forgetting factor is used to identify and continuously update the time-varying stiffness parameters and time-varying damping parameters in the dynamic model online.

[0035] The dynamics of a single joint are simplified to a second-order linear system, whose continuous-time differential equations are expressed as: ; in The equivalent rotational inertia of the joint is considered constant and is determined by the robotic arm's structural design and obtained through offline calibration. , and These are the joint angle, angular velocity, and angular acceleration, respectively. and The time-varying damping coefficient and time-varying stiffness coefficient are to be identified, both of which drift slowly with the aging degree of the composite material and changes in ambient temperature and humidity. This is the measured driving torque; This represents the composite disturbance moment resulting from the unmodeled dynamics, wind load, and load variation. To facilitate online identification, the above equations are discretized. Let the sampling period be... Typically, the value is taken as 1 millisecond, then the angular acceleration can be approximated as... After rearranging the equation, we get: ; Ignore high-frequency disturbance terms The dominant influence in parameter identification (due to its rate of change being much smaller than parameter drift) is treated as slowly varying noise, and a linear regression model is constructed accordingly: ; The output = Regression vector The parameter vector to be identified Using a forgetting factor The recursive least squares algorithm for Perform recursive estimation. During algorithm initialization, set initial parameter estimates. The nominal value is the covariance matrix. For example, a large diagonal matrix This is to reflect the initial uncertainty. In each control cycle k, the following sub-steps are executed sequentially: Calculate the gain vector: ; Update parameter estimates: ; Update the covariance matrix: ; Forgetting factor The value of is strictly limited to between 0.98 and 0.995. This range ensures that the algorithm can effectively track the slow drift of parameters without causing drastic fluctuations in the estimated value due to excessive forgetting of historical data. The identified... and As the core input for subsequent disturbance observer and controller gain calculations, its real-time performance and accuracy directly determine the upper limit of the entire adaptive control system's performance.

[0036] Then, step S3 is executed, whereby a nonlinear disturbance observer for estimating the composite disturbance moment is constructed based on the online identified time-varying stiffness and damping parameters. The design goal of this observer is to estimate the composite disturbance moment. Separate from and estimate in real time from system dynamics. Define auxiliary state variables. Its dynamic equation is: This equation constructs a high-gain observation structure by introducing the observer gain L(t), thus enabling... It can quickly converge to the internal state associated with the disturbance. The estimated value of the composite disturbance moment. It can be calculated using the following formula: ; Observer Gain A diagonal matrix designed to be positive definite has a unique non-zero element. satisfy ,in It is a constant greater than 0, typically ranging from 5 to 10.

[0037] This design makes the observer bandwidth proportional to the square root of the system stiffness, so that when the joint stiffness increases due to temperature rise... When the bandwidth decreases, the observer gain automatically decreases to avoid amplifying measurement noise due to excessive bandwidth; conversely, when the stiffness increases, the gain increases to maintain a sufficient disturbance estimation speed.

[0038] This adaptive gain mechanism ensures that the disturbance observer has good robustness and convergence under different operating conditions.

[0039] Observer state The numerical integrator updates in real time, and the integration step size is synchronized with the main control cycle.

[0040] To prevent integral saturation, for Set reasonable upper and lower limits, which are determined based on the statistical analysis of the historical maximum disturbance amplitude.

[0041] Next, step S4 is executed, where the nonlinear disturbance observer is used to calculate in real time the estimated value of the composite disturbance torque acting on each joint, and a feedforward compensation torque of equal magnitude but opposite direction to the estimated value is generated. This feedforward compensation torque... Defined as .

[0042] This torque is directly added to the final control command to actively counteract the effects of external disturbances on the system.

[0043] Since the disturbance observer has made a unified estimate of wind load, load abrupt changes and unmodeled dynamics, the compensation mechanism does not need to model various disturbance sources separately and has strong generalization ability.

[0044] The calculation of the compensation torque is completed within each control cycle, with a delay of less than one sampling cycle, ensuring the timeliness of the compensation.

[0045] Then, step S5 is executed, whereby the proportional gain, integral gain, and derivative gain of the proportional-integral-derivative main controller are dynamically and in real time calculated and adjusted based on the time-varying stiffness parameters and time-varying damping parameters identified online and according to the preset closed-loop system pole configuration strategy.

[0046] The closed-loop system is expected to have standard second-order dynamic characteristics, with the characteristic equation being: Damping ratio The value is fixed at 0.707, which corresponds to the optimal balance point between critical damping and underdamping, enabling the fastest response with no or minimal overshoot; undamped natural frequency. The value is dynamically selected based on the current task requirements, ranging from 10 to 30 radians per second. For delicate operations, such as wire stripping or bolt tightening, Use a lower value (e.g., 12 radians per second) to ensure smoothness; for large-scale transfer tasks, Choose a higher value (e.g., 25 radians per second) to improve efficiency.

[0047] Based on this expected dynamic, the controller gain is calculated in real time according to the following formula: Proportional gain: ; Differential gain: ; Integral gain: ; Where α is the integral gain adjustment coefficient, and its value ranges from 0.1 to 0.3.

[0048] This gain calculation formula originates from matching the actual system dynamics with the desired closed-loop dynamics, ensuring that regardless of... and No matter how things change, the closed-loop poles always remain at the preset positions. The expression introduced and The nonlinear relationship aims to maintain the coordination between integral action and proportional-differential action, and to prevent integral saturation or excessive phase lag.

[0049] All gain values ​​are recalculated based on the latest identified parameters in each control cycle and applied to the controller immediately.

[0050] Then, step S6 is executed: based on the desired joint motion trajectory and combined with the time-varying stiffness and damping parameters identified online, the theoretical feedforward torque used to counteract the joint's own dynamics is calculated. Desired trajectory Provided by the upper-level task planning module, along with its first derivative. With the second derivative .

[0051] Theoretical feedforward torque The calculation formula is: ; This torque is generated entirely based on the currently identified dynamic model and is used to apply the torque required in advance to overcome joint inertia, damping, and elasticity, thereby significantly reducing the burden on the feedback controller. The acceleration signal is obtained by smoothing the discrete trajectory points through cubic spline interpolation and then analytically differentiating it, ensuring that the acceleration signal is continuous and free of high-frequency oscillations.

[0052] Finally, step S7 is executed, in which the feedback control torque calculated by the proportional-integral-derivative main controller based on the trajectory tracking error, the theoretical feedforward torque, and the feedforward compensation torque are algebraically summed to generate the final synthetic control torque applied to the joint drive unit.

[0053] Tracking error Defined as Feedback control torque Calculated using the standard PID formula: ; The integral term is achieved through numerical integration using the trapezoidal rule, and the differential term is achieved through integration with respect to the trapezoidal rule. The first-order difference is performed and then low-pass filtered to obtain the final synthesized control torque. ; The torque command is sent to the joint servo driver, which drives the motor to output the corresponding torque, enabling the joint to accurately track the desired trajectory. The entire control loop operates in a high-speed real-time operating system, with the control cycle strictly locked at 1 millisecond to ensure data synchronization and computational timeliness between modules.

[0054] At the system level, the adaptive control system for the insulated robotic arm used in live-line working platforms includes a multi-source sensor module, a joint drive unit, a real-time computing unit, and a communication bus network.

[0055] The multi-source sensor module integrates the aforementioned multi-turn absolute encoder, strain gauge torque sensor, platinum resistance temperature sensor, capacitive humidity sensor, and three-dimensional ultrasonic anemometer. All sensors are connected to the analog / digital input interface of the real-time computing unit via shielded twisted-pair cables. The sampling clock is synchronized by the same crystal oscillator to eliminate phase errors caused by asynchronous sampling.

[0056] The joint drive unit includes a servo motor, a reducer, and a built-in current loop controller. It receives torque commands from the real-time calculation unit and feeds back actual current and status information via the CANopen bus.

[0057] The real-time computing unit employs a multi-core processor architecture, with one core dedicated to executing the aforementioned adaptive control algorithm, while the remaining cores handle communication, logging, and security monitoring. The communication bus network uses the industrial Ethernet protocol to ensure low-latency transmission of control commands and status data.

[0058] After the system is powered on, it first performs sensor self-test and zero-point calibration, then enters the parameter identification and initialization stage. After the identified parameters converge to a reasonable range, the disturbance observer and adaptive controller are started, and the system enters the normal operation mode.

[0059] Throughout the operation, the system continuously monitors the tracking error, control torque amplitude, and parameter identification residual of each joint. If any indicator exceeds the preset threshold, a degraded control strategy or a safe shutdown is immediately triggered to ensure the safety of live-line work.

[0060] In summary, this embodiment constructs a composite adaptive control system capable of simultaneously coping with time-varying internal parameters and strong external disturbances by deeply integrating three core technologies: online parameter identification, adaptive disturbance observation, and dynamic gain tuning.

[0061] This system not only solves the fundamental defects of traditional PID control in the application of insulated robotic arms, but also improves the control accuracy and stability to a new level through multiple feedforward compensation mechanisms, providing a solid technical guarantee for the safe and efficient execution of high-voltage live-line work.

[0062] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0063] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An adaptive control method for an insulated robotic arm of a live working platform, characterized in that, The method comprises the following steps: Real-time acquisition of the state information of each joint of the insulated mechanical arm, wherein the joint state information comprises joint angle, joint angular velocity, joint driving torque, joint body temperature, ambient humidity around the joint, and wind speed vector of the position where the live working platform is located; A second-order lumped parameter dynamic model of a single joint of the insulated mechanical arm is established, and a recursive least squares algorithm with a forgetting factor is used to identify and continuously update the time-varying stiffness parameters and time-varying damping parameters in the dynamic model according to the real-time acquired joint state information; Based on the identified time-varying stiffness parameters and time-varying damping parameters, a nonlinear disturbance observer for estimating the composite disturbance torque is constructed, wherein the composite disturbance torque comprises equivalent disturbance torque caused by wind load, end load change and model unmodeled dynamics; The nonlinear disturbance observer is used to calculate the estimated value of the composite disturbance torque acting on each joint in real time, and a feedforward compensation torque equal in size and opposite in direction to the estimated value is generated; According to the identified time-varying stiffness parameters and time-varying damping parameters, and according to the preset closed-loop system pole placement strategy, the proportional gain, integral gain and differential gain of the proportional-integral-differential main controller are dynamically and real-timely calculated and adjusted; Based on the desired joint motion trajectory, the time-varying stiffness parameters and time-varying damping parameters are combined to calculate a theoretical feedforward torque for offsetting the joint dynamics; The feedback control torque calculated by the proportional-integral-differential main controller based on the trajectory tracking error, the theoretical feedforward torque and the feedforward compensation torque are algebraically summed to generate a final composite control torque applied to the joint driving unit, so as to drive the insulated mechanical arm joint to accurately track the desired joint motion trajectory.

2. The adaptive control method for the insulating mechanical arm of the hot-line work platform according to claim 1, wherein, The real-time acquisition of the state information of each joint of the insulated mechanical arm specifically comprises: The real-time angle of the joint is acquired through the multi-turn absolute value encoder installed at the output end of each joint; The real-time angular velocity of the joint is acquired by performing first-order differential operation on the joint angle signal and filtering through a low-pass digital filter; The real-time driving torque applied to the joint is acquired through the strain torque sensor arranged between the joint driving motor and the reducer; The real-time temperature of the joint body is acquired through the platinum resistance temperature sensor closely attached to the surface of the joint composite material shell; The ambient humidity around the joint is acquired through the capacitive humidity sensor arranged at the base of the mechanical arm; A three-dimensional ultrasonic anemometer is installed on the top of the live working platform to acquire the three-dimensional wind speed vector at the position of the platform, and the wind speed vector is calculated in the coordinate system of each joint according to the real-time configuration of the mechanical arm.

3. The adaptive control method for the insulated robotic arm of the live working platform according to claim 1, wherein, The step of establishing a second-order lumped parameter dynamic model of a single joint of the insulated mechanical arm and identifying online by using a recursive least squares algorithm with a forgetting factor specifically comprises: The dynamic behavior of a single joint is described by the following differential equation: ； wherein is the moment of inertia of the joint, , is the angle of the joint, is the angular velocity of the joint, and is the angular acceleration of the joint, are the time-varying damping parameter and the time-varying stiffness parameter to be identified, is the driving torque, is the compound disturbance torque; The above differential equation is transformed into a linear regression form in discrete time: ； is a sampling period; regression vector ; parameter vector to be identified ; The recursive least square algorithm with a forgetting factor is used to update the parameter vector The update law is ； ； ； wherein is the recognized parameter estimate of the time instant, is a gain matrix, is a covariance matrix, is a forgetting factor.

4. The adaptive control method for the insulated robotic arm of the live working platform according to claim 1, wherein, The nonlinear disturbance observer for estimating the composite disturbance torque specifically comprises: Defining state variables The dynamic equation is: ； wherein is an observer gain, and are time-varying damping and stiffness parameter estimates output by the online identification module; the estimated value of the compound disturbance torque is calculated by the following equation: ； The observer gain is designed as a positive definite diagonal matrix whose diagonal element values are proportional to the square root of the time-varying stiffness parameter to ensure that the observer bandwidth can be adaptively adjusted with the change of system stiffness.

5. The adaptive control method for the insulated robotic arm of the live working platform according to claim 1, wherein, The gain parameters of the proportional-integral-derivative main controller are dynamically and real-timely calculated and adjusted according to a preset closed-loop system pole placement strategy, and the specific steps include: The characteristic equation of the desired closed-loop control system is set to have a second-order standard form: ； wherein is the desired damping ratio, is the desired undamped natural frequency; the proportional gain of the proportional-integral-derivative master controller the integral gain and the derivative gain in real time according to the following relationship: ； ； ； wherein is a gain adjustment factor.

6. The adaptive control method for the insulated robotic arm of a live working platform according to claim 1, wherein, The steps of calculating the theoretical feedforward torque and generating the final synthesized control torque include: Acquiring a desired articulation trajectory and its first derivative and second derivative ; The theoretical feedforward torque is calculated: ； The trajectory tracking error is calculated: ； The feedback control torque of the proportional-integral-derivative main controller is calculated: ； The feedforward compensation torque is generated: ； The final synthesized control torque is calculated: ； And the torque command is sent to the joint servo driver.

7. The adaptive control method for the insulated robotic arm of the live working platform according to claim 6, wherein, a trajectory tracking error a first derivative obtained by first differencing the and filtering through a low pass digital filter.

8. The adaptive control method for the insulated robotic arm of the live working platform according to claim 3, wherein, Moment of inertia of the joint Obtained by offline calibration and considered constant during control.

9. The adaptive control method for the insulated robotic arm of the live working platform according to claim 2, wherein, The process of resolving the three-dimensional wind speed vector into each joint coordinate system includes: according to the pose vector composed of the current joint angles of the robot arm, the global wind speed vector is converted into each joint local coordinate system through a homogeneous transformation matrix.

10. The adaptive control method for the insulated robotic arm of the live working platform according to claim 4, wherein, State variable of the nonlinear disturbance observer Updated in real time by a numerical integrator, the integration step is synchronized with the main control period and is Upper and lower limits are set to prevent integral saturation.

Images