CN-121424360-B - Flying robot control method, device and medium based on force feedback

CN121424360BCN 121424360 BCN121424360 BCN 121424360BCN-121424360-B

Abstract

The application provides a flying robot control method, equipment and medium based on force feedback, wherein the method comprises the steps of deploying a six-dimensional force moment sensor on an end effector of flying robot operation, and acquiring and screening effective moment data in real time to obtain a force error signal; the compensation layer based on Lyapunov self-adaptive law is utilized, the total disturbance is estimated in a cooperative mode by combining a dual observer, and the disturbance is injected into the gesture controller as a feedforward quantity; and then calculating the adjustment quantity according to the target impedance model and the force error, and generating an executing mechanism control instruction. The technical scheme provided by the application can realize stable control of aerial interactive operation, and effectively improve the disturbance rejection capability and interactive force control precision of the flying robot.

Inventors

WANG JIANMING
JI YAFEI

Assignees

北京飞舆科技有限公司

Dates

Publication Date: 20260512
Application Date: 20251104

Claims (8)

1. The flying robot control method based on force feedback is characterized by comprising the following steps: S1, acquiring moment data of a flying robot through a six-dimensional force moment sensor arranged on an end effector of the flying robot, filtering and screening the moment data to acquire effective moment data, and comparing the effective moment data with preset moment data to acquire a force error signal; S2, estimating total disturbance of the flying robot in real time through a compensation layer based on a Lyapunov self-adaptive law, injecting the total disturbance into a gesture controller as a feedforward quantity, calculating adjustment quantity data required for realizing expected force interaction according to a target impedance model and an actual measurement force error, inputting the adjustment quantity data into the gesture controller enhanced by feedforward compensation, and generating a control instruction for driving an executing mechanism of the flying robot, wherein the total disturbance comprises effective moment data and external environment disturbance data; Constructing an operation mechanical state feature vector, wherein the feature vector comprises triaxial orthogonal force data, triaxial torque data, a flight attitude angle, an angular speed and position information; The LSTM-Kalman filtering fusion prediction model learns time sequence association characteristics of the mechanical state feature vector through an LSTM network and performs error correction on a prediction result by combining with Kalman filtering; the predicted information of the environmental disturbance is fed forward to a local path planner, and an anti-disturbance optimized track is generated on line by combining the stability constraint of the working force and the flight stability index; Converting the optimized track into a control instruction, and sending the control instruction to a control system of the flying robot for execution; through the compensation layer based on Lyapunov self-adaptive law, the total disturbance of the flying robot is estimated in real time, and the method specifically comprises the following steps: Constructing a dual observer cooperative estimation architecture, wherein the dual observer cooperative estimation architecture comprises a nonlinear disturbance observer based on a dynamic model and a parameter adaptive device based on a Lyapunov adaptive law; The nonlinear disturbance observer is used for rapidly estimating lumped disturbance caused by an external environment and unmodeled dynamics; the parameter self-adaption device is used for identifying and compensating slow-change systematic disturbance caused by mechanical arm movement, load change and model parameter uncertainty on line; and fusing the lumped disturbance estimated by the nonlinear disturbance observer with the systematic disturbance identified by the parameter self-adaptive device to obtain the equivalent total disturbance of the flying robot.
2. The method of claim 1, wherein the acquiring moment data of the flying robot, and the filtering and screening the moment data to acquire effective moment data specifically comprises: on the working end effector of the flying robot, a six-dimensional force moment sensor is deployed to acquire the interaction between the flying robot and an operator in real time Triaxial orthogonal force And torque about three axes ; Performing low-pass filtering and noise reduction on the acquired triaxial orthogonal force data and triaxial torque data; Converting the processed triaxial orthogonal force data and triaxial torque data from a sensor coordinate system to a body coordinate system or a world coordinate system of the flying robot through a coordinate transformation algorithm; the method for acquiring effective moment data by identifying the effective contact state based on the preset contact force threshold specifically comprises the following steps: and when the force value of at least one direction of the converted triaxial orthogonal force data and triaxial torque data exceeds a preset threshold value and the duration is more than or equal to 2 sampling periods of the six-dimensional force moment sensor, judging that the three-axis orthogonal force data and the triaxial torque data are in effective contact and activating a feedback control loop, and taking the triaxial orthogonal force data and the triaxial torque data which are corresponding at the moment as the effective force moment data.
3. The method of claim 1, wherein the external environmental disturbance data is obtained by online estimation by an adaptive gain dynamic disturbance observer, and specifically comprises online estimating, by the adaptive gain dynamic disturbance observer, unmodeled aerodynamic disturbances caused by high wind speeds, turbulence, humidity changes, airflow vortices and sudden gusts and the influence thereof on attitude angles, wherein gain coefficients of the observer are dynamically adjusted according to absolute values of disturbance estimation errors, gain is increased when the errors are increased to increase response speed, and gain is reduced when the errors are reduced to suppress noise.
4. The method of claim 1, wherein the target impedance model is expressed as: ; Wherein M, D, K are respectively set virtual inertia, virtual damping and virtual stiffness matrices, Is the deviation of the position of the object, Is a force error signal.
5. The method of claim 1, further comprising, after outputting the predicted information on the environmental disturbance: The disturbance intensity level of the prediction information adopts a hierarchical adjustment strategy to dynamically adjust a desired force set value, a target impedance parameter and a gain parameter of a self-adaptive compensation algorithm, wherein the disturbance intensity level is divided based on the influence amplitude of environmental disturbance on the interaction force deviation of the flying robot and an operator, and the division is based on the stability requirement of the interaction operation of the matched flying robot, the low disturbance level is a level that the interaction force deviation occupies a preset interaction force threshold value proportion to be in a first preset interval, the medium disturbance level is a level that the deviation proportion is in a second preset interval, the high disturbance level is a level that the deviation proportion is in a third preset interval, the first preset interval upper limit is less than the second preset interval lower limit, the second preset interval upper limit is less than the third preset interval lower limit, the first preset interval is an interval that the deviation proportion is lower than the minimum disturbance influence threshold value allowed by the operation stability, the second preset interval is an interval that the deviation proportion is between the minimum disturbance influence threshold value and the maximum acceptable disturbance influence threshold value, and the third preset interval is an interval that the deviation proportion is higher than the maximum acceptable disturbance influence threshold value; when the disturbance level is low, the adjustment amplitude of each parameter is based on the condition that the control parameter is not caused to oscillate and the stability of the interaction between the flying robot and the operation object is maintained; when the disturbance level is medium, the adjustment amplitude of each parameter is based on the balanced disturbance response speed and the control stability, and the adjustment amplitude is larger than the adjustment amplitude corresponding to the low disturbance level; when the disturbance level is high, the adjustment amplitude of each parameter is higher than the adjustment amplitude corresponding to the medium disturbance level so as to ensure that the disturbance is restrained rapidly and avoid the expansion of the interaction force deviation.
6. The method according to claim 1, wherein the method further comprises: recording sensor data, control instructions and state variables in real time, wherein the state variables comprise angular acceleration, force change rate and environmental air pressure value of the flying robot; The system comprises a performance evaluation module, a power tracking error calculation unit, an attitude stability index calculation unit and a power analysis unit, wherein the performance evaluation module comprises a data preprocessing unit, an error calculation unit and an index analysis unit, the data preprocessing unit carries out noise reduction and standardization processing on recorded data, the error calculation unit calculates the power tracking error based on the processed data, and the index analysis unit calculates the attitude stability index in combination with a flight attitude fluctuation range; The method comprises the steps of adjusting control parameters or an optimization model according to an evaluation result, wherein the method specifically comprises the steps of fine-adjusting the control parameters on line when a force tracking error is in a preset small error interval and gesture stability index fluctuation is in a preset small fluctuation interval, and optimizing the model off line when the force tracking error exceeds the preset small error interval or the gesture stability index fluctuation exceeds the preset small fluctuation interval.
7. An electronic device, comprising: Processor, and A memory arranged to store computer executable instructions that when executed cause the processor to perform the steps of the force feedback based flying robot control method according to any one of claims 1-6.
8. A storage medium storing computer executable instructions which when executed implement the steps of the force feedback based flying robot control method of any one of claims 1-6.

Description

Flying robot control method, device and medium based on force feedback Technical Field The present document relates to the field of flying robot control technologies, and in particular, to a method, apparatus, and medium for controlling a flying robot based on force feedback. Background With the development of artificial intelligence and robot technology, flying robots are widely applied in the field of aloft work by virtue of maneuverability and flexibility, and cover electric power facility maintenance, wind power generation equipment maintenance, aloft structure installation and detection, disaster monitoring and emergency response. The application scene has common technical difficulties of complex operation environment, inaccessible operation position and extremely high requirements on operation stability and precision. The flying robot needs to complete accurate physical interaction with the environment or an operator, such as contact, grabbing, pushing and the like, sense and adapt to interaction force and reaction force while maintaining stable flying posture, and provides serious challenges for a stability control technology. The existing flying robot control multi-dependence high-precision positioning navigation technology realizes position and track maintenance, combines path planning to avoid static obstacles, but has obvious defects in air interaction operation that firstly dynamic interaction force cannot be effectively processed, six-dimensional force/torque disturbance generated by contact of a mechanical arm and an operator is poor in robustness, secondly the flying robot is sensitive to external environment disturbance, gesture balance is easy to damage such as high wind speed and turbulence, quick compensation is difficult, thirdly an effective force feedback mechanism is lacking, force sensor data is not integrated into a gesture control loop in a closed loop mode, fourthly the control strategy fusion degree is low, advanced force control strategy and gesture self-adaptive compensation algorithm are not deep, and thirdly high-dimensional mechanical information of the six-dimensional force/torque sensor is not fully utilized, and a high-efficiency chemical state adjustment mechanism is difficult to construct. Aiming at the problems, the related patent proposes partial solutions, but has obvious limitations that, for example, although the patent CN111984024B realizes robust tracking of position and posture through geometric control, a real-time force feedback mechanism is not established, a six-dimensional force/torque sensor is not utilized to sense interaction force, a force control strategy is not integrated, and the patent CN112527008B simplifies a control structure through decoupling design, but does not involve force feedback and mechanical state adjustment, and does not combine an advanced force control strategy. In summary, the prior art fails to solve the core problem, namely how to integrate multidimensional force sensing, attitude self-adaptive compensation and advanced force control strategies through a high-efficiency real-time force feedback mechanism, so as to realize real-time accurate adjustment of the aerial interaction operation mechanical state of the flying robot, and the flying attitude stability, interaction operation precision and safety under a complex environment are difficult to guarantee synchronously. Disclosure of Invention The invention provides a flying robot control method, a device and a medium based on force feedback, and aims to solve the problems. According to an embodiment of the present invention, there is provided a flying robot control method based on force feedback, including: S1, acquiring moment data of a flying robot through a six-dimensional force moment sensor arranged on an end effector of the flying robot, filtering and screening the moment data to acquire effective moment data, and comparing the effective moment data with preset moment data to acquire a force error signal; S2, estimating total disturbance of the flying robot in real time through a compensation layer based on a Lyapunov self-adaptive law, injecting the total disturbance into a gesture controller as a feedforward quantity, calculating adjustment quantity data required for realizing expected force interaction according to a target impedance model and an actual measurement force error, inputting the adjustment quantity data into the gesture controller enhanced by feedforward compensation, and generating a control instruction for driving an executing mechanism of the flying robot, wherein the total disturbance comprises effective moment data and external environment disturbance data. According to an embodiment of the present invention, there is provided an electronic apparatus including: Processor, and A memory arranged to store computer executable instructions that, when executed, cause the processor to perform the steps of a flying robot control method based on force feedback a