A method and system for impact force mitigation based on power detection

Abstract

The present disclosure provides a method and system for impact force suppression based on power detection. The method detects the contact force received by the end of the robot actuator, obtains the contact force change x according to the contact force, generates the impact force suppression compensation of the end of the actuator according to the contact force change x, and compensates the expected motion trajectory. Compared with the commonly used compliant control strategy, a different control strategy is adopted, the "dynamic" power is used as the input of the controller, a different power compensation module is designed, the synchronous adjustment is carried out through the forward input channel of the force closed loop controller, the impact vibration can be reduced, the larger contact force at the end of the robot can be suppressed, and the stability of the robot operation can be improved.

Description

Technical Field

[0001] The present invention relates to the field of robotics technology, and in particular to a method and system for suppressing impact force of a robot terminal mechanism. Background Art

[0002] Robots play an important role in intelligent manufacturing, deep space exploration, disaster relief, assisted medical care, home services and other fields. The joint drive mechanisms are the "skeleton and muscles" of robots, and their high-performance control is the core foundation for robots to complete their tasks.

[0003] Linear hydraulic actuators are widely used in large-load robots due to their high power-to-weight ratio, simple structure, and lack of intermediate reduction mechanisms. The output force of the linear hydraulic actuator is converted into an output force at the end of the robot through the mechanical structure, which then acts on the "environmental load." This "environmental load" can be the object being manipulated or the surrounding environment.

[0004] To complete specific tasks, a robot's end-point (such as a robotic arm's gripper or a biomimetic robot's hand or foot) must move along a desired trajectory. During this motion, the end-point undergoes both a free motion phase (without contact with environmental loads) and a constrained motion phase (with contact with environmental loads). In unpredictable environments, the transition between free and constrained motion can generate impact vibrations between the motion mechanism and the environmental loads. This impact vibration can cause the robot's motion to be jerky, impacting its operational performance and stability.

[0005] In an unknown contact environment, the position of the environmental load has a certain degree of uncertainty, and it is difficult for the controller to accurately estimate the contact time and position between the end of the motion mechanism and the load. This results in the pre-planned motion trajectory being unable to adapt to the actual motion situation. At this time, the driving force calculated based on the dynamic model will undergo unpredictable mutations, resulting in a large step in the reference force output, which seriously affects the stability and compliance of the motion mechanism.

[0006] like Figure 1 As shown in the figure, the unpredictable contact between the end of the motion mechanism and the environmental load can be divided into three stages: initial contact stage, transition stage, and continuous contact stage.

[0007] In the initial contact stage, the motion trajectory of the motion mechanism produces a sudden change in speed due to the obstruction of the environmental load. According to the principle of conservation of energy, ,in, For impact force, is the impact force acting time, is the mass of the object, is the change in the object's velocity, so the magnitude of the impact force is Regarding the generation of impact force, a preliminary analysis has been conducted in Research Aspect 1. The magnitude of this force depends on the duration of this phase and is unrelated to the maximum output capacity of the motion mechanism. The duration of this phase, in turn, is closely related to the material and structure of the "mechanism-environment" interface.

[0008] During the transition phase, the motion mechanism and the environmental load themselves will absorb part of the impact force through a certain deformation, so the contact force will decrease.

[0009] During the continuous contact stage, it can be divided into two situations: one is that the base of the motion mechanism and the environmental load is fixed. At this time, if the motion mechanism continues to move according to the originally set expected motion trajectory, the contact force between the two will continue to increase as shown by the solid line part in the figure until the output capacity of the motion mechanism is reached; the other is that the base of the motion mechanism or the environmental load is movable. At this time, relative motion occurs between the two under the action of the contact force, and the contact force is maintained at a low level.

[0010] From the preliminary analysis above, it can be seen that implementing the impact force suppression strategy before the transition stage is the most appropriate time to reduce impact vibration.

[0011] In order to reduce the impact vibration between the robot end and the environmental load, there are usually the following methods: the conventional passive vibration reduction and compliance method relying on springs, dampers, etc. can no longer meet the needs, and a special active compliance control strategy needs to be designed.

[0012] After the impact force is generated, it can be equivalently considered that the output force of the motion mechanism deviates from the expected value. There are two ways to correct this deviation: one is through the feedback path of the force closed-loop controller, and the other is through the forward input channel of the force closed-loop controller. The first path is the controller's inherent adjustment capability, but the feedback path alone cannot effectively solve the problem of the continuous increase in contact force during the sustained contact phase. Therefore, the present invention uses the second path, namely, the forward input channel of the force closed-loop controller, to perform synchronous adjustment.

[0013] For the second approach, a commonly used active compliance control strategy is impedance / admittance control. In impedance control, the difference between the desired and actual positions is fed into an impedance controller. The controller uses the position difference and preset impedance parameters to calculate a force deviation. This force deviation is then added to the desired force to generate a final force signal, which serves as input to the force controller, causing the robot to exhibit impedance characteristics.

[0014] In contrast, admittance control employs the opposite input-output logic. In admittance control, a force deviation signal is fed into an admittance controller, which uses the deviation signal and specific admittance parameters to calculate the deviation between the desired and actual positions. This calculated deviation is then added to the desired position to form the input command to the position controller, causing the robot to exhibit admittance characteristics.

[0015] The above two commonly used compliant control strategies use the end force value as the controller input, which will produce a continuous large position deviation when the external force changes slightly or remains unchanged. Summary of the Invention

[0016] In view of this, the present invention provides an impact force suppression method based on dynamic detection. Compared with the commonly used compliant control strategy, it adopts a different control strategy, uses "dynamic" force as the controller input, and designs a different dynamic compensation module. Through the forward input channel of the force closed-loop controller, synchronous adjustment is performed, which can reduce impact vibration, suppress the large contact force at the end of the robot, and improve the stability of the robot operation.

[0017] In order to solve the above technical problems, the present invention is implemented as follows.

[0018] A method for suppressing impact force based on dynamic detection, comprising:

[0019] Detect the contact force on the end of the robot actuator;

[0020] Obtain contact force variation x according to the contact force;

[0021] Generate the impact force suppression compensation amount at the end of the actuator based on the contact force change x , and compensate for the desired motion trajectory.

[0022] Preferably, the contact force includes three-dimensional force and three-dimensional moment.

[0023] Preferably, obtaining the contact force variation according to the contact force is as follows: obtaining the contact force variation by differential method and filtering, and the transfer function is expressed as:

[0024] (1)

[0025] in, is a first-order differential operator with filtering effect, s is the Laplace operator, is the filtering effect adjustment coefficient.

[0026] Preferably, the impact force suppression compensation amount of the actuator end is generated according to the contact force variation. Using formula (2):

[0027] (2)

[0028] in, is the compensation gain adjustment parameter; x is the contact force change after processing by formula (1), is a nonlinear mapping function about x, used to map the contact force variation x into a position correction; The function satisfies the requirements of symmetry, non-decreasing and boundedness.

[0029] Preferably, the nonlinear mapping function is an inverse tangent function or a hyperbolic tangent function.

[0030] The present invention also provides an impact force suppression system based on power detection, comprising a closed-loop system consisting of a kinematic model, an actuator position controller, an actuator force controller, and an actuator mechanism. The impact force suppression system also includes a multi-dimensional force sensor and a power compensation feedback channel.

[0031] The multi-dimensional force sensor is used to detect the contact force applied to the end of the robot's actuator and send it to the power compensation module;

[0032] The power compensation feedback channel includes a power detection module and a compensation generation module;

[0033] The power detection module is used to obtain the contact force variation x according to the contact force and send it to the compensation generation module;

[0034] The compensation generation module is used to generate the impact force suppression compensation amount of the actuator end according to the contact force change x , which is fed forward to the closed-loop system to compensate for the desired actuator end motion trajectory.

[0035] Preferably, the multi-dimensional force sensor collects the three-dimensional force and three-dimensional torque at the end of the actuator and sends them to the power compensation module for six-dimensional force processing.

[0036] Preferably, the power detection module obtains the contact force variation by differential method and performs filtering, and the transfer function is expressed as:

[0037] (1)

[0038] in, is a first-order differential operator with filtering effect, s is the Laplace operator, is the filtering effect adjustment coefficient.

[0039] Preferably, the compensation generation module generates the impact force suppression compensation amount at the end of the actuator using formula (2):

[0040] (2)

[0041] in, is the compensation gain adjustment parameter; x is the contact force change after processing by formula (1), is a nonlinear mapping function about x, used to map the contact force variation x into a position correction; The function satisfies the requirements of symmetry, non-decreasing and boundedness.

[0042] Preferably, the nonlinear mapping function is an inverse tangent function or a hyperbolic tangent function.

[0043] Beneficial effects:

[0044] (1) The present invention does not directly calculate the compensation amount based on the end force value. Instead, it obtains the contact force change x, i.e., the dynamic force, also referred to as the "dynamic" force in the present invention, based on the force detection results. Based on the "dynamic" force detection results, the force is adjusted through the forward input of the force closed-loop controller. The use of "dynamic" force can avoid the significant end position error caused by small contact force changes but large contact forces, and can also make position compensation for contact force changes, thereby effectively reducing the impact vibration between the robot end and the environmental load.

[0045] (2) The present invention calculates the compensation amount based on the "dynamic" force. According to the characteristics of the "dynamic" force, the compensation calculation function is selected to meet the requirements of symmetry, non-reducibility and boundedness. When the end is affected by the rapidly changing contact force, the present invention can generate a compensation amount that is positively correlated with the amplitude of the "dynamic" force according to the positive and negative fluctuations of the "dynamic" force, and the positive and negative values ​​are the same as the "dynamic" force, and the amplitude is bounded, thereby effectively suppressing the vibration of the end and avoiding the generation of oscillations.

[0046] (3) In a specific embodiment, an inverse tangent function or a hyperbolic tangent function is selected as the compensation calculation function, which is simple to calculate and easy to implement.

[0047] (4) In a specific embodiment, filtering the differentiated detection force can suppress interference and spikes, further improving the impact vibration suppression effect between the robot end and the environmental load.

[0048] (5) The suppression strategy of the present invention belongs to feedback control. On the basis of high real-time performance, it does not rely on the accuracy of modeling and has stronger applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 Schematic diagram of the impact contact process.

[0050] Figure 2 Schematic diagram of the impact force suppression system based on dynamic detection of the present invention. DETAILED DESCRIPTION

[0051] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0052] The present invention provides a solution for suppressing the impact force at the end of a robot actuator. The basic idea is to calculate the contact force variation x, i.e., the dynamic change of the contact force, referred to as "dynamic" force, based on the detected contact force at the end of the robot actuator. Then, the "dynamic" force compensation strategy is used to generate the impact force suppression compensation at the end of the actuator. , the desired motion trajectory of the position loop is corrected in real time, and the compensation output generated by the position loop controller corrects the input channel of the force controller.

[0053] The specific implementation steps of the present invention are as follows:

[0054] Step 1: Detect the contact force on the end of the robot actuator.

[0055] In this step, a multi-dimensional sensor can be installed at the end of the robot actuator to measure the contact force on the end of the actuator, including three-dimensional force and three-dimensional torque.

[0056] Step 2: Dynamic detection: Obtain the contact force change x based on the contact force.

[0057] In this step, the contact force is converted into the contact force variation x, which is also the force in the above text and the accompanying figure. , and suppress low-frequency interference and errors to obtain the "dynamic" force of the end. The power detection part of this step can be constructed as follows:

[0058] (1)

[0059] From the transfer function, we can see that It is actually a first-order differential operator with a high-pass filtering effect, which suppresses the low-frequency components in the end force value, so that the high-frequency components, that is, the "dynamic" force part, can pass through almost without attenuation, realizing the function of extracting the end "dynamic" force data. In addition, it can also filter the received end force, suppress the low-frequency noise and zero bias error of the force sensor, and thus reduce the total error. The parameter It is the filter effect adjustment coefficient, used to adjust the cutoff frequency; is the Laplace operator.

[0060] Step 3: Compensation generation: Generate the impact force suppression compensation at the end of the actuator based on the contact force change x , and compensate for the desired motion trajectory.

[0061] The compensation generation operation in this step adopts formula (2)

[0062] (2)

[0063] Among them, the compensation generation part can generate a position correction value according to the "dynamic" force, thereby generating a compensation displacement .in, It is the compensation gain adjustment parameter, which is used to adjust the amplification factor of the compensation amount; is a nonlinear mapping function about x, which can map the dynamic change of force x into the correction of position .

[0064] Functions for power design Requires certain requirements to be met:

[0065] First, due to the change in force (i.e. x) can be positive or negative. When the absolute values ​​are equal and the signs are opposite, the compensation amount should also be equal in absolute value and opposite in sign, so The symmetry needs to be satisfied, that is ;

[0066] Second, in When the value increases, the compensation amount should also increase or remain unchanged, so It needs to satisfy the non-decreasing property, that is ;

[0067] Third, in When changing, the function output value should have upper and lower limits, otherwise it may cause the end position to change too much or cause oscillation, so Need to satisfy boundedness;

[0068] In summary, function The requirements of symmetry, non-subtraction and boundedness must be met. In one embodiment of the present invention, the nonlinear function is selected is the inverse tangent function , or the hyperbolic tangent function , both can achieve good compensation effect. a is the coefficient in the function.

[0069] The power generated by the environment is collected in real time through power detection, and the compensation generation part composed of the above nonlinear function is input to generate compensation displacement in real time. , thereby compensating the end force and allowing the end actuator to maintain a more accurate motion state under environmental interference and impact.

[0070] Based on the above impact force suppression method, the present invention also provides an impact force suppression system based on power detection, which includes a traditional closed-loop system, a multi-dimensional force sensor, and a power compensation feedback channel.

[0071] like Figure 2 As shown in Figure 1, the closed-loop system includes a kinematic model, an actuator position controller, an actuator force controller, and an actuator. The power compensation feedback channel includes a power detection module and a compensation generation module.

[0072] The kinematic model is used to calculate the corresponding motion trajectory of the actuator based on the end reference trajectory of the actuator through kinematic modeling.

[0073] The actuator position controller is used to control the position of the actuator. It uses a given reference position to decouple the reference force under constraint conditions and calculates the required actuator reference force.

[0074] The actuator force controller is used to control the torque of the actuator. According to the input actuator reference force, the output force of the actuator follows the reference force to drive the actuator.

[0075] The actuator follows the end reference track Perform exercises to achieve corresponding functions.

[0076] The system input is the reference trajectory of the actuator end After processing the system's kinematic model, the end actuator reference trajectory is determined. This reference trajectory is then fed into the position controller, which then outputs the actuator reference force. After decoupling, this serves as the actuator force controller, thereby driving the actuator's motion. However, changes in external loads can affect the actuator's following performance, causing disturbances in the actual trajectory. Therefore, a dynamic compensation feedback channel is added to the original closed-loop control system. The control system's given trajectory is adjusted based on the output of the dynamic compensation module, thereby reducing the impact of external disturbances.

[0077] The multi-dimensional force sensor is used to detect the contact force applied to the end of the robot's actuator in real time and send it to the power compensation module.

[0078] The function of the dynamic compensation feedback channel is to calculate the dynamic change based on the output force of the actuator detected by the multi-dimensional force sensor, and convert the change x into the compensation input of the position loop. , for the desired motion trajectory Perform compensation and obtain the compensated motion trajectory .

[0079] The power compensation feedback channel includes a power detection module and a compensation generation module.

[0080] The dynamic detection module is used to obtain the contact force variation x based on the contact force and send it to the compensation generation module. In one embodiment, the dynamic detection module uses a differential method to obtain the contact force variation and performs filtering. The transfer function is expressed as the above formula (1).

[0081] The compensation generation module is used to generate the impact force suppression compensation at the end of the actuator according to the contact force change x , which is fed forward to the closed-loop system to compensate for the desired actuator end motion trajectory. In one embodiment, the compensation generation module generates the suppression compensation using formula (2). The nonlinear mapping function in the compensation algorithm can be an inverse tangent function or a hyperbolic tangent function.

[0082] The above specific embodiments merely illustrate the design principles of the present invention. The shapes and names of the components described herein may vary and are not limiting. Therefore, those skilled in the art may modify or substitute equivalents for the technical solutions described in the above embodiments. Such modifications and substitutions, without departing from the inventive spirit and technical solutions of the present invention, shall fall within the scope of protection of the present invention.

Claims

1. A method for suppressing impact force based on dynamic detection, characterized in that: include: Detect the contact force on the end of the robot actuator; Obtain contact force variation based on contact force: Use differential method to obtain contact force variation and perform filtering. The transfer function is expressed as: （1） in, is a first-order differential operator with filtering effect, s is the Laplace operator, is the filter effect adjustment coefficient; x is the change of contact force after processing by formula (1); According to the contact force variation x Generates the impact force suppression compensation amount at the end of the actuator , to compensate for the desired motion trajectory; where, is the compensation gain adjustment parameter; For about x The nonlinear mapping function is used to convert the contact force variation x A correction amount mapped to a position; The function meets the requirements of symmetry, non-decreasing and boundedness. Choose the inverse tangent function or the hyperbolic tangent function.

2. The method according to claim 1, wherein The contact force includes three-dimensional force and three-dimensional moment.

3. An impact force suppression system based on dynamic detection, comprising a closed-loop system consisting of a kinematic model, an actuator position controller, an actuator force controller, and an actuator, characterized in that: The impact force suppression system also includes a multi-dimensional force sensor and a power compensation feedback channel; The multi-dimensional force sensor is used to detect the contact force applied to the end of the robot's actuator and send it to the power compensation module; The power compensation feedback channel includes a power detection module and a compensation generation module; The power detection module is used to obtain the contact force variation according to the contact force and send it to the compensation generation module. The power detection module obtains the contact force variation by differential method and performs filtering. The transfer function is expressed as: （1） in, is a first-order differential operator with filtering effect, s is the Laplace operator, is the filter effect adjustment coefficient; x is the change of contact force after processing by formula (1); The compensation generation module is used to generate the compensation according to the contact force variation. x Generates the impact force suppression compensation amount at the end of the actuator , which is fed forward to the closed-loop system to compensate for the desired actuator end motion trajectory; The impact force suppression compensation amount is generated using formula (2): （2） in, is the compensation gain adjustment parameter; For about x The nonlinear mapping function is used to convert the contact force variation x A correction amount mapped to a position; The function meets the requirements of symmetry, non-decreasing and boundedness. Choose the inverse tangent function or the hyperbolic tangent function.

4. The impact force suppression system according to claim 3, wherein: The multi-dimensional force sensor collects the three-dimensional force and three-dimensional torque at the end of the actuator and sends them to the power compensation module for six-dimensional force processing.

Images