CN117182929A - Flexible control method and device for on-orbit assembly of double-arm robot - Google Patents
Flexible control method and device for on-orbit assembly of double-arm robot Download PDFInfo
- Publication number
- CN117182929A CN117182929A CN202311461575.9A CN202311461575A CN117182929A CN 117182929 A CN117182929 A CN 117182929A CN 202311461575 A CN202311461575 A CN 202311461575A CN 117182929 A CN117182929 A CN 117182929A
- Authority
- CN
- China
- Prior art keywords
- target object
- double
- model
- arm
- current
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 72
- 238000013528 artificial neural network Methods 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 16
- 230000007613 environmental effect Effects 0.000 claims description 22
- 230000008569 process Effects 0.000 claims description 21
- 230000009471 action Effects 0.000 claims description 18
- 239000013598 vector Substances 0.000 claims description 17
- 239000012636 effector Substances 0.000 claims description 15
- 230000002787 reinforcement Effects 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 11
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000013016 damping Methods 0.000 claims description 8
- 230000009466 transformation Effects 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 claims description 6
- 239000003795 chemical substances by application Substances 0.000 claims description 6
- 230000009977 dual effect Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000003993 interaction Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000001360 synchronised effect Effects 0.000 description 3
- 230000001133 acceleration Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002427 irreversible effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- HOWHQWFXSLOJEF-MGZLOUMQSA-N systemin Chemical compound NCCCC[C@H](N)C(=O)N[C@@H](CCSC)C(=O)N[C@@H](CCC(N)=O)C(=O)N[C@@H]([C@@H](C)O)C(=O)N[C@@H](CC(O)=O)C(=O)OC(=O)[C@@H]1CCCN1C(=O)[C@H]1N(C(=O)[C@H](CC(O)=O)NC(=O)[C@H](CCCN=C(N)N)NC(=O)[C@H](CCCCN)NC(=O)[C@H](CO)NC(=O)[C@H]2N(CCC2)C(=O)[C@H]2N(CCC2)C(=O)[C@H](CCCCN)NC(=O)[C@H](CO)NC(=O)[C@H](CCC(N)=O)NC(=O)[C@@H](NC(=O)[C@H](C)N)C(C)C)CCC1 HOWHQWFXSLOJEF-MGZLOUMQSA-N 0.000 description 1
- 108010050014 systemin Proteins 0.000 description 1
Abstract
The invention relates to the technical field of space double-arm robot compliance control, in particular to a compliance control method and device for on-orbit assembly of a double-arm robot. The method comprises the following steps: acquiring the current motion state of the double-arm robot; inputting the motion state of the double-arm robot into a pre-trained target model to obtain the expected track of the target object at the current moment and the control parameters of an impedance control model which is suitable for the current environment; the target model is obtained by training a preset neural network by taking the motion state of the double mechanical arms, the operation force state of the double mechanical arms and the motion state of a target object as training samples; based on the expected track of the target object, the control parameters of the impedance control model and the preset double-ring impedance control model, the expected joint angle of the double-arm robot which is suitable for the current environment is obtained to be used as a control instruction of the double-arm robot, so that the flexible control of the double-arm robot is realized. The invention can improve the efficiency and flexibility of the on-orbit assembly of the double-arm robot.
Description
Technical Field
The invention relates to the technical field of space double-arm robot compliance control, in particular to a compliance control method and device for on-orbit assembly of a double-arm robot.
Background
With the wide application of ground robot technology and the rapid development of aerospace technology, the application of robots to space on-orbit services has demonstrated great advantages and benefits. Compared with a single-arm robot system, the cooperative operation of the double-arm robot can process more diversified operation tasks, and meanwhile, the system has strong load capacity and is suitable for operation of large-inertia objects or flexible objects, so that the application of the double-arm robot to space on-orbit service has important significance.
At present, simple motion planning and position control cannot realize double-arm cooperative operation related to force-position coordination, so that a flexible control method is needed to ensure that a target object of the double-arm robot does not slide off or cause irreversible damage to a system in the whole on-orbit service process.
In the related art, the compliant control methods all adopt frames separated by a motion planning and an impedance controller, and when the position of a target object changes, a path is often required to be re-planned, so that the on-orbit assembly efficiency is low; in addition, the parameters of the impedance controller in the framework remain substantially unchanged during operation, and the adaptability and robustness of the algorithm are poor, so that the flexibility of double-arm operation is poor.
Based on this, a method and a device for flexibly controlling on-orbit assembly of a double-arm robot are needed to solve the technical problems of low efficiency and poor flexibility of the double-arm robot during on-orbit assembly.
Disclosure of Invention
In order to solve the technical problems of low efficiency and poor flexibility of on-orbit assembly of the double-arm robot, the embodiment of the invention provides a flexible control method and device for on-orbit assembly of the double-arm robot.
In a first aspect, embodiments of the present disclosure provide a compliant control method for on-orbit assembly of a dual-arm robot, including:
acquiring the current motion state of the double-arm robot;
inputting the current motion state of the double-arm robot into a pre-trained target model to obtain an expected track of a current target object and control parameters of an impedance control model which is suitable for the current environment; the target model is obtained by training a preset neural network by taking the motion state of the double mechanical arms, the operation force state of the double mechanical arms and the motion state of a target object as training samples;
and obtaining an expected joint angle of the double-arm robot which is suitable for the current environment based on the expected track of the current target object, the control parameters of the impedance control model and the preset double-loop impedance control model, and taking the expected joint angle as a control instruction of the double-arm robot to realize the flexible control of the double-arm robot.
In a second aspect, an embodiment of the present invention further provides a compliant control apparatus for on-orbit assembly of a dual-arm robot, including:
the acquisition module is used for acquiring the current motion state of the double-arm robot;
the input module is used for inputting the current motion state of the double-arm robot into a pre-trained target model so as to obtain the expected track of the current target object and the control parameters of the impedance control model which are suitable for the current environment;
the calculation module is used for obtaining the expected joint angle of the double mechanical arm which is suitable for the current environment based on the expected track of the current target object, the control parameters of the impedance control model and the preset double-ring impedance control model so as to realize the flexible control of the double-arm robot.
In a third aspect, embodiments of the present specification further provide an electronic device, including a memory and a processor, where the memory stores a computer program, and the processor implements the method described in any embodiment of the present specification when executing the computer program.
In a fourth aspect, the embodiments of the present specification also provide a computer-readable storage medium having stored thereon a computer program which, when executed in a computer, causes the computer to perform the method according to any of the embodiments of the present specification.
The embodiment of the specification provides a method and a device for flexibly controlling on-orbit assembly of a double-arm robot, which are characterized in that the obtained current motion state of the double-arm robot is input into a pre-trained target model to obtain a desired track of a current target object and control parameters of an impedance model suitable for the current environment, and finally, the desired joint angle of the double-arm robot suitable for the current environment is obtained based on the desired track of the current target object, the control parameters of the impedance model and a preset double-loop impedance control model to serve as a control instruction of the double-arm robot, so that flexible control of the double-arm robot is realized. Therefore, the scheme can realize the synchronous optimization of the motion planning of the double mechanical arms of the robot and the control parameters of the impedance control model, and can adaptively adjust the parameters of the impedance controller according to the current environmental characteristics, thereby improving the on-orbit assembly efficiency of the double-arm robot and the operation flexibility of the double-arm robot.
Drawings
In order to more clearly illustrate the embodiments of the present description or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, and it is obvious that the drawings in the following description are some embodiments of the present description, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a compliance control method for on-orbit assembly of a dual-arm robot in accordance with one embodiment of the present disclosure;
FIG. 2 is a hardware architecture diagram of an electronic device according to an embodiment of the present disclosure;
FIG. 3 is a block diagram of a compliant control apparatus for on-track assembly of a dual arm robot in accordance with one embodiment of the present disclosure.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present specification more apparent, the technical solutions of the embodiments of the present specification will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present specification, and it is apparent that the described embodiments are some, but not all, embodiments of the present specification, and all other embodiments obtained by persons of ordinary skill in the art without making any inventive effort based on the embodiments of the present specification are within the scope of protection of the present specification.
Referring to fig. 1, an embodiment of the present disclosure provides a compliant control method for on-orbit assembly of a dual-arm robot, the method including:
step 100: acquiring the current motion state of the double-arm robot;
step 102: inputting the current motion state of the double-arm robot into a pre-trained target model to obtain an expected track of a current target object and control parameters of an impedance control model which is suitable for the current environment; the target model is obtained by training a preset neural network by taking the motion state of the double mechanical arms, the operation force state of the double mechanical arms and the motion state of a target object as training samples;
step 104: based on the expected track of the current target object, the control parameters of the impedance control model and the preset double-ring impedance control model, the expected joint angle of the double-arm robot which is suitable for the current environment is obtained to be used as a control instruction of the double-arm robot, so that the flexible control of the double-arm robot is realized.
In this embodiment, the obtained current motion state of the dual-arm robot is input into a pre-trained target model to obtain a desired track of a target object and control parameters of an impedance model adapted to a current environment, and finally, based on the desired track of the target object, the control parameters of the impedance model and a preset dual-loop impedance control model, a desired joint angle of the dual-arm robot adapted to the current environment is obtained to be used as a control instruction of the dual-arm robot, so that flexible control of the dual-arm robot is realized. Therefore, the scheme can realize the motion planning of the double mechanical arms of the robot and the synchronous optimization of the control parameters of the impedance control model, and can adaptively adjust the parameters of the impedance controller according to the current environmental characteristics, thereby improving the on-orbit assembly efficiency of the double-arm robot and the operation flexibility of the double-arm robot.
The manner in which the individual steps shown in fig. 1 are performed is described below.
For step 100, the current motion state of the dual-arm robot is obtained.
In this embodiment, the motion state of the dual-arm robot at the current time specifically refers to the pose and the speed of the end effector of the dual-arm robot at the current time.
For step 102:
in one embodiment of the present disclosure, the motion state of the dual-robot includes a joint angle, a joint angular velocity of the dual-robot, a pose and a velocity of the dual-robot end effector;
the operation force state of the double mechanical arms comprises six-dimensional force information of the tail ends of the double mechanical arms and environmental external force born by the mass center of the target object;
the motion state of the target object includes the pose of the centroid of the target object.
In the implementation, in order to obtain the expected motion of the target object at the current moment and the parameter adjustment strategy of the impedance controller which is adaptive to the current environment, the obtained motion state of the double mechanical arms, the operation force information of the double mechanical arms and the motion state of the target object are taken as training samples, and the on-line training of the motion planning of the target object and the parameter adjustment strategy of the impedance controller is carried out through a preset neural network until the algorithm converges; in order to improve learning efficiency, an error vector between the current position of the object and the target position is also used as state information of the training sample.
In one embodiment of the present disclosure, the preset neural network includes an LSTM network module, an MLP network module, a deep reinforcement learning network module, and an experience playback pool module connected in sequence.
Considering that the deep reinforcement learning network algorithm has the characteristic of strong generalization, the autonomy of system planning and control can be improved, and in the embodiment, based on the deep reinforcement learning network, an LSTM network, an MLP network, a deep reinforcement learning network and an experience playback pool are designed into a centralized neural network structure, so that the self-adaptive adjustment of control parameters of a dual-mechanical arm motion planning and impedance control model is unified under the same framework, and the problems of low on-orbit assembly efficiency of the dual-mechanical arm caused by separating the mechanical arm motion planning and the impedance control model in the related technology and poor flexibility of the dual-mechanical arm caused by poor adaptability and robustness of a traditional impedance control model are solved.
In one embodiment of the present description, the object model is trained by:
acquiring an environmental state quantity according to a preset frequency; the environment state quantity comprises an operation force state of the double mechanical arms, a motion state of the double mechanical arms and a motion state of a target object;
inputting the operation force state of the double mechanical arms into an LSTM network module to obtain the operation force state of the double mechanical arms with time sequences;
performing feature extraction on the operation force state of the double mechanical arms by using the MLP network module to obtain a feature vector of the double mechanical arm force state at the current moment;
taking the characteristic vector of the double mechanical arm strength state, the motion state of the double mechanical arm and the motion state of the target object as input state vectors of the deep reinforcement learning network module;
setting an environmental reward function of the intelligent agent, and training the intelligent agent by using a deep reinforcement learning network module to obtain an output action of a target object; the output actions of the target object comprise the position variation of the mass center of the target object and the control parameters of the impedance control model;
storing the obtained environmental state quantity, the output action of the target object, the environmental rewarding value obtained by executing the output action and the environmental state quantity after executing the action into an experience playback pool module to obtain experience data;
updating the current neural network by using the experience data until the algorithm enters a convergence state, and obtaining a target model.
In the embodiment, the motion state of the double mechanical arms and the motion state of the target object are obtained according to the preset frequency of 20Hz, and the stability of the double mechanical arms is guaranteed under the frequency; meanwhile, assuming that the control frequency of the impedance control model is h, presetting a time sequence of the operating force state of the double mechanical arms with the length of h/20, inputting the time sequence of the operating force state of the double mechanical arms with the shape of [ h/20,3] into an LSTM network module at each moment t, and outputting the time sequence of the operating force state with the same shape by the LSTM network module (wherein 3 represents that the operating force state of the double mechanical arms is the force in the three-dimensional direction); then, a layer of MLP network module is adopted to conduct feature extraction on the time sequence of the operation force state output by the LSTM network module, and a feature vector of the operation force state at the current moment is obtained; and finally, connecting the motion state of the mechanical arm, the motion state of the target object and the characteristic vector of the operation force state at the current moment, which are obtained at the current moment, so that the motion state and the motion state of the target object are used as the input state vector of the deep reinforcement learning network.
In this embodiment, considering that the dual-arm of the dual-arm robot has stronger timing when cooperatively operating the same target object, but the conventional reinforcement learning network cannot process the problem of high-dimensional timing information, a centralized neural network structure with an LSTM network module and an MLP network module is designed, and the centralized neural network structure can synchronously acquire the motion state and the external interaction force state of the dual-arm, so as to realize the motion planning of the dual-arm of the robot and the synchronous optimization of the parameters of the impedance control model, so that the dual-arm robot can adaptively adjust the control parameters of the impedance control model according to the current environmental characteristics and the motion state of the system in the operation process, and further improve the on-orbit assembly efficiency and the flexibility of the dual-arm operation of the dual-arm robot.
Meanwhile, in the embodiment, the LSTM network module is adopted to acquire the high-dimensional time sequence force state information of the double mechanical arms acquired at each moment, and then the MLP network module is adopted to extract the characteristics of the force state information of the time sequence output by the LSTM network module, so that the flexible operation performance is obviously improved due to the fact that the characteristic vector has the strong trend characteristic.
For step 104:
in one embodiment of the present disclosure, the step 104 may specifically include the following steps:
obtaining an expected track and an expected operating force of the tail end of the double mechanical arms according to the expected track of the target object and the kinematic constraint type and dynamic model of the closed-chain system;
obtaining a flexible expected track of the tail end of the double mechanical arm according to control parameters of the impedance control model, a preset double-ring control model and the expected track and expected operation force of the tail end of the double mechanical arm;
and obtaining the expected joint angle of the double mechanical arms according to the flexible expected track at the tail ends of the double mechanical arms and the inverse kinematics formula of the mechanical arms.
In this embodiment, the target model obtained after training is applied to the cooperative assembly operation of the two mechanical arms of the robot, and the trained target model obtains the position variation of the centroid of the current target object (i.e., the expected track of the target object) and the control parameters of the impedance control model adapted to the current environment by obtaining the motion state of the two mechanical arms of the robot at the current moment. In the impedance control model, the desired inertia has a large influence on the stability of the whole dual-arm robot system, but has a small influence on the final motion state of the dual-arm robot, and the desired inertia is set to a fixed parameter, for example, may be 1 during training and during the whole on-orbit assembly. Therefore, the control parameters of the impedance control model output by the target model in this embodiment are damping parameters and stiffness parameters adapted to the current environment.
In one embodiment of the present disclosure, the preset dual-loop impedance control model includes an outer loop impedance model and an inner loop impedance model; the outer loop impedance model is used for correcting the expected track of the target object, and the inner loop impedance model is used for correcting the expected track of the double mechanical arms;
first, a second-order impedance model in Cartesian space is established:
in the method, in the process of the invention,、/>and->Desired inertia, damping and stiffness of the second order impedance model, respectively,>、/>and->The actual acceleration, speed and position of the manipulator in cartesian space, +.>、/>And->Desired acceleration, speed and position of the arm, respectively +.>For the actual contact force +.>For the expected contact force of the mechanical arm, three impedance parameters in the embodiment are positive definite matrixes, and diagonal matrixes are generally selected to obtain linear decoupling response;
an outer ring impedance model is established between a target object and an external environment, and a second-order impedance model in a Cartesian space is rewritten to obtain an outer ring impedance control model:
，/>，/>
in the method, in the process of the invention,is the actual trajectory of the target object +.>And desired track->Track error between->Is the first derivative of the target object trajectory error, < >>Is the second derivative of the target object trajectory error, +.>Is the external force that the environment exerts on the target object, < >>、/>And->Desired inertia, damping and stiffness of the outer loop impedance model, respectively,>for the actual trajectory of the target object->First derivative of>For the desired trajectory of the target object->First derivative of>For the actual trajectory of the target object->Second derivative of>Is the object of
Actual trajectory of objectIs a second derivative of (2);
then, an inner loop impedance control model of the tail end operation force of the mechanical arm i is established, wherein the inner loop impedance model is as follows:
，/>，/>，/>
in the method, in the process of the invention,is the actual trajectory of the arm i +.>And desired track->Track error between->Arm i distal desired force->And actual operating force->Error between->、/>And->Desired inertia, damping and stiffness of the inner loop impedance model, respectively,>is the actual trajectory of the arm i +.>First derivative of>For the desired trajectory of the arm i +.>First derivative of>Is the actual trajectory of the arm i +.>Second derivative of>For the desired trajectory of the arm i +.>Is of the second order of (2)And (3) derivative.
In this embodiment, the preset outer loop impedance control model is used to correct the expected track of the current target object, when no interaction occurs between the target object and the environment, the interference force of the environment on the target object is minimized, the outer loop impedance model does not correct the expected track of the target object any more, and the expected track of the target object is obtained, so that the compliance control between the target object and the environment is realized.
In one embodiment of the present disclosure, in the cooperative motion process of the dual mechanical arm, the dual mechanical arm and the target object need to be always kept in stable and rigid connection, so that the actuator at the end of the dual mechanical arm and the target object need to satisfy the closed-chain kinematic constraint at all times, and the kinematic constraint formula of the closed-chain system is as follows:
in the method, in the process of the invention,is a homogeneous transformation matrix of the end effector coordinate system of the mechanical arm i relative to the base coordinate system thereof,/>For homogeneous transformation matrix of world coordinate system relative to mass center coordinate system of target object>The pose of the contact point between the end effector of the arm i and the target object is obtained,/->The base coordinate system of the mechanical arm i is inverted to obtain the base coordinate system;
the dynamic model comprises a force balance equation and a moment balance equation of the target object; the force balance equation of the target object is as follows:
the moment balance equation of the target object is:
in the method, in the process of the invention,and->External force and moment applied to the target object by the environment, respectively,/->,/>,/>And->Force and moment applied to the target object by the left and right mechanical arm end effectors, respectively, +.>,/>，/>Are respectively->,，/>Vector of force on target object acting on object centroid,/->And->The linear and angular velocities, respectively, of the centroid of the target object,/->Is the mass of the target object, < > and->Moment of inertia of the target object->Is the weight of the target object.
In this embodiment, first, the above-mentioned kinematic constraint of the closed-chain system is adopted to decompose the expected compliant track of the current target object, specifically, first, the transformation matrix of the centroid of the target object relative to the world coordinate system at each moment is obtained from the compliant track of the target objectThe inverse of this matrix is found and substituted into the kinematic constraint of the closed-chain systemIn the method, a transformation matrix of the mechanical arm end effector relative to a base coordinate system is calculated>The expected track of the double-arm end effector can be obtained, and the expected operating force of the double-arm end is calculated according to the dynamic model; substituting the control parameters of the impedance control model which is output by the target model and is suitable for the current environment, and the expected track and the expected operating force of the tail end of the double mechanical arm which are obtained through calculation into the inner ring impedance control model, and obtaining the flexible track of the tail end of the double mechanical arm through calculation; and finally, calculating an inverse kinematics formula of the mechanical arm to obtain the expected joint angle of the double mechanical arms corresponding to the flexible track, wherein the expected joint angle is used as a control instruction of the double mechanical arms, so that flexible control of on-orbit assembly of the robot is realized.
In summary, the embodiment provides a compliant control method and device for on-orbit assembly of a dual-arm robot, which unifies motion planning of a target object and adjustment of parameters of an impedance control model by using a centralized neural network structure, and the centralized neural network structure can synchronously acquire a motion state and an interaction force state of the dual-arm, so that control parameters of the impedance control model can be adaptively adjusted according to the current environmental characteristics and the motion state of a system, and further, the on-orbit assembly efficiency of the dual-arm and the operation flexibility of the dual-arm are improved. Meanwhile, the LSTM network module and the MLP network module are adopted to conduct feature extraction on the force state information in operation in advance, and the flexible operation performance is remarkably improved due to force trend features in the feature vectors.
As shown in fig. 2 and 3, the embodiment of the present disclosure provides a compliant control apparatus for on-track assembly of a dual-arm robot. The apparatus embodiments may be implemented by software, or may be implemented by hardware or a combination of hardware and software. In terms of hardware, as shown in fig. 2, a hardware architecture diagram of an electronic device where a compliant control apparatus for on-orbit assembly of a dual-arm robot is provided in the embodiments of the present disclosure is shown, where the electronic device where the embodiment is located may generally include other hardware, such as a forwarding chip responsible for processing a message, in addition to the processor, the memory, the network interface, and the nonvolatile memory shown in fig. 2. Taking a software implementation as an example, as shown in fig. 3, the device in a logic sense is formed by reading a corresponding computer program in a nonvolatile memory into a memory by a CPU of an electronic device where the device is located and running the computer program.
As shown in fig. 3, the present embodiment provides a compliant control apparatus for on-orbit assembly of a dual-arm robot, including:
an obtaining module 300, configured to obtain a current motion state of the dual-arm robot;
the input module 302 is configured to input a motion state of the dual-arm robot into a pre-trained target model, so as to obtain a desired track of a target object and control parameters of an impedance control model adapted to a current environment;
the calculation module 304 is configured to obtain, based on the desired trajectory of the target object, the control parameters of the impedance control model, and the preset dual-loop impedance control model, a desired joint angle of the dual-arm robot adapted to the current environment, so as to implement compliant control of the dual-arm robot.
In the embodiment of the present disclosure, the obtaining module 300 may be used to perform the step 100 in the embodiment of the method, the input module 302 may be used to perform the step 102 in the embodiment of the method, and the calculating module 304 may be used to perform the step 104 in the embodiment of the method.
In one embodiment of the present disclosure, the motion state of the dual-mechanical arm includes a joint angle, a joint angular velocity of the dual-mechanical arm, a pose and a velocity of the dual-mechanical arm end effector;
the operating force state of the double mechanical arms comprises six-dimensional force information of the tail ends of the double mechanical arms and environmental external force applied to the mass center of the target object;
the motion state of the target object comprises the pose of the mass center of the target object.
In one embodiment of the present disclosure, the preset neural network includes an LSTM network module, an MLP network module, a deep reinforcement learning network module, and an experience playback pool module that are sequentially connected.
In one embodiment of the present specification, the object model is trained by:
acquiring an environmental state quantity according to a preset frequency; the environment state quantity comprises an operation force state of the double mechanical arms, a motion state of the double mechanical arms and a motion state of a target object;
inputting the operating force state of the double mechanical arms into the LSTM network module to obtain the operating force state of the double mechanical arms with time sequences;
extracting the characteristics of the operation force state of the double mechanical arms by using the MLP network module to obtain the characteristic vector of the double mechanical arm force state at the current moment;
taking the characteristic vector of the double mechanical arm strength state, the motion state of the double mechanical arm and the motion state of the target object as input state vectors of the deep reinforcement learning network module;
setting an environmental reward function of the intelligent agent, and training the intelligent agent by using the deep reinforcement learning network module to obtain an output action of a target object; the output action of the target object comprises the position variation of the mass center of the target object and control parameters of an impedance control model;
storing the obtained environmental state quantity, the output action of the target object, the environmental rewarding value obtained by executing the action and the environmental state quantity after executing the action into the experience playback pool module to obtain experience data;
and updating the current neural network by using the experience data until the algorithm enters a convergence state, so as to obtain the target model.
In one embodiment of the present specification, the preset dual-loop impedance model includes an outer-loop impedance model and an inner-loop impedance model; wherein,
the outer loop impedance model is:
，/>，/>
in the method, in the process of the invention,is the track error between the actual track and the desired track of the target object,/or->Is the first derivative of the target object trajectory error, < >>Is the object ofSecond derivative of body track error,/>Is the external force that the environment exerts on the target object, < >>、/>And->The desired inertia, damping and stiffness of the outer loop impedance model, respectively;
the inner ring impedance model is:
，/>，/>
in the method, in the process of the invention,is the track error between the actual track and the desired track of the arm i, < >>Error between the desired force and the actual operating force of the end of the arm i>、/>And->Respectively inner ringsThe desired inertia, damping and stiffness of the impedance model.
In one embodiment of the present specification, the computing module is configured to perform the following operations:
obtaining an expected track and an expected operating force of the tail end of the double mechanical arms according to the expected track of the target object and the kinematic constraint and dynamic model of the closed-chain system;
obtaining a flexible expected track of the tail end of the double mechanical arm according to the control parameters of the impedance control model, the preset double-ring control model and the expected track and expected operation force of the tail end of the double mechanical arm;
and obtaining the expected joint angle of the double mechanical arms according to the expected flexible track at the tail ends of the double mechanical arms and the inverse kinematics formula of the mechanical arms.
In one embodiment of the present specification, the kinematic constraint formula of the closed-chain system is:
in the method, in the process of the invention,is a homogeneous transformation matrix of the end effector coordinate system of the mechanical arm i relative to the base coordinate system thereof,/>For homogeneous transformation matrix of world coordinate system relative to mass center coordinate system of target object>The pose of the contact point between the end effector of the arm i and the target object is obtained,/->The base coordinate system of the mechanical arm i is inverted to obtain the base coordinate system;
the dynamic model comprises a force balance equation and a moment balance equation of the target object; wherein, the force balance equation of the target object is:
the moment balance equation of the target object is as follows:
in the method, in the process of the invention,and->External force and moment applied to the target object by the environment, respectively,/->,/>,/>And->Force and moment exerted on the target object by the end effectors of the left and right mechanical arms, respectively +.>，/>，/>Are respectively->,/>，At the object of interestVector of force on body to centroid of object, +.>And->The linear and angular velocities, respectively, of the centroid of the target object,/->Is the mass of the target object, < > and->Moment of inertia of the target object->Is the weight of the target object.
It will be appreciated that the structure illustrated in the embodiments of the present disclosure is not intended to be limiting in detail with respect to a compliant control apparatus for an on-track assembly of a dual arm robot. In other embodiments of the present description, a evasive maneuver control apparatus for a spacecraft may include more or fewer components than illustrated, or may combine certain components, or split certain components, or may be a different arrangement of components. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.
The content of information interaction and execution process between the modules in the above-mentioned device, because the content is based on the same conception as the method embodiment of the present specification, the specific content can be referred to the description in the method embodiment of the present specification, and the description is not repeated here.
The embodiment of the specification also provides electronic equipment, which comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the processor realizes the flexible control method for the on-orbit assembly of the double-arm robot in any embodiment of the specification.
The embodiments of the present specification also provide a computer readable storage medium having a computer program stored thereon, which when executed by a processor, causes the processor to perform a compliance control method for on-track assembly of a two-arm robot in any of the embodiments of the present specification.
Specifically, a system or apparatus provided with a storage medium on which a software program code realizing the functions of any of the above embodiments is stored, and a computer (or CPU or MPU) of the system or apparatus may be caused to read out and execute the program code stored in the storage medium.
In this case, the program code itself read from the storage medium may realize the functions of any of the above-described embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present specification.
Examples of the storage medium for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (e.g., CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a nonvolatile memory card, and a ROM. Alternatively, the program code may be downloaded from a server computer by a communication network.
Further, it should be apparent that the functions of any of the above-described embodiments may be implemented not only by executing the program code read out by the computer, but also by causing an operating system or the like operating on the computer to perform part or all of the actual operations based on the instructions of the program code.
Further, it is understood that the program code read out by the storage medium is written into a memory provided in an expansion board inserted into a computer or into a memory provided in an expansion module connected to the computer, and then a CPU or the like mounted on the expansion board or the expansion module is caused to perform part and all of actual operations based on instructions of the program code, thereby realizing the functions of any of the above embodiments.
It is noted that relational terms such as first and second, and the like, are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the above method embodiments may be implemented by hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, where the program, when executed, performs steps including the above method embodiments; and the aforementioned storage medium includes: various media in which program code may be stored, such as ROM, RAM, magnetic or optical disks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present specification, and are not limiting thereof; although the present specification has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present specification.
Claims (10)
1. The method for flexibly controlling the on-orbit assembly of the double-arm robot is characterized by comprising the following steps of:
acquiring the current motion state of the double-arm robot;
inputting the current motion state of the double-arm robot into a pre-trained target model to obtain an expected track of a current target object and control parameters of an impedance control model which is suitable for the current environment; the target model is obtained by training a preset neural network by taking the motion state of the double mechanical arms, the operation force state of the double mechanical arms and the motion state of a target object as training samples;
and obtaining an expected joint angle of the double-arm robot which is suitable for the current environment based on the expected track of the current target object, the control parameters of the impedance control model and the preset double-loop impedance control model, and taking the expected joint angle as a control instruction of the double-arm robot to realize the flexible control of the double-arm robot.
2. The method of claim 1, wherein the motion state of the dual-robot arm comprises joint angle, joint angular velocity, pose and velocity of the dual-robot arm end effector;
the operating force state of the double mechanical arms comprises six-dimensional force information of the tail ends of the double mechanical arms and environmental external force applied to the mass center of the target object;
the motion state of the target object comprises the pose of the mass center of the target object.
3. The method of claim 1, wherein the predetermined neural network comprises an LSTM network module, an MLP network module, a deep reinforcement learning network module, and an empirical playback pool module connected in sequence.
4. A method according to claim 3, wherein the target model is trained by:
acquiring an environmental state quantity according to a preset frequency; the environment state quantity comprises an operation force state of the double mechanical arms, a motion state of the double mechanical arms and a motion state of a target object;
inputting the operating force state of the double mechanical arms into the LSTM network module to obtain the operating force state of the double mechanical arms with time sequences;
extracting the characteristics of the operation force state of the double mechanical arms by using the MLP network module to obtain the characteristic vector of the double mechanical arm force state at the current moment;
taking the characteristic vector of the double mechanical arm strength state, the motion state of the double mechanical arm and the motion state of the target object as input state vectors of the deep reinforcement learning network module;
setting an environmental reward function of the intelligent agent, and training the intelligent agent by using the deep reinforcement learning network module to obtain an output action of a target object; the output action of the target object comprises the position variation of the mass center of the target object and control parameters of an impedance control model;
storing the obtained environmental state quantity, the output action of the target object, the environmental rewarding value obtained by executing the action and the environmental state quantity after executing the action into the experience playback pool module to obtain experience data;
and updating the current neural network by using the experience data until the algorithm enters a convergence state, so as to obtain the target model.
5. The method of claim 1, wherein the pre-set dual loop impedance model comprises an outer loop impedance model and an inner loop impedance model; wherein, the outer loop impedance model is:
，/>，/>
in the method, in the process of the invention,is the track error between the actual track and the desired track of the target object,/or->Is the first derivative of the target object trajectory error, < >>Is the second derivative of the target object trajectory error, +.>Is the external force that the environment exerts on the target object, < >>、/>And->The desired inertia, damping and stiffness of the outer loop impedance model, respectively;
the inner ring impedance model is:
，/>，/>
in the method, in the process of the invention,is the track error between the actual track and the desired track of the arm i, < >>Error between the desired force and the actual operating force of the end of the arm i>、/>And->The desired inertia, damping and stiffness of the inner loop impedance model, respectively.
6. The method according to claim 1, wherein the obtaining the desired joint angle of the dual-mechanical arm adapted to the current environment based on the desired trajectory of the current target object, the control parameters of the impedance control model, and the preset dual-loop impedance control model includes:
obtaining an expected track and an expected operating force of the tail end of the double mechanical arms according to the expected track of the current target object, the kinematic constraint type and the dynamic model of the closed-chain system;
obtaining a flexible expected track of the tail end of the double mechanical arm according to the control parameters of the impedance control model, the preset double-ring control model and the expected track and expected operation force of the tail end of the double mechanical arm;
and obtaining the expected joint angle of the double mechanical arms according to the expected flexible track at the tail ends of the double mechanical arms and the inverse kinematics formula of the mechanical arms.
7. The method of claim 6, wherein the step of providing the first layer comprises,
the kinematic constraint of the closed chain system is as follows:
in the method, in the process of the invention,is a homogeneous transformation matrix of the end effector coordinate system of the mechanical arm i relative to the base coordinate system thereof,/>For homogeneous transformation matrix of world coordinate system relative to mass center coordinate system of target object>The pose of the contact point between the end effector of the arm i and the target object is obtained,/->The base coordinate system of the mechanical arm i is inverted to obtain the base coordinate system;
the dynamic model comprises a force balance equation and a moment balance equation of the target object; wherein, the force balance equation of the target object is:
the moment balance equation of the target object is as follows:
in the method, in the process of the invention,and->External force and moment applied to the target object by the environment, respectively,/->, />,/>And->Force and moment applied to the target object by the left and right mechanical arm end effectors, respectively, +.>,/>，/>Are respectively->,/>，/>Vector of force on target object acting on object centroid,/->And->The linear and angular velocities, respectively, of the centroid of the target object,/->Is the mass of the target object, < > and->Moment of inertia of the target object->Is the weight of the target object.
8. A compliant control apparatus for on-orbit assembly of a dual-arm robot, comprising:
the acquisition module is used for acquiring the current motion state of the double-arm robot;
the input module is used for inputting the current motion state of the double-arm robot into a pre-trained target model so as to obtain the expected track of the target object at the current moment and the control parameters of the impedance control model which are suitable for the current environment;
the calculation module is used for obtaining the expected joint angle of the double mechanical arms which is suitable for the current environment based on the expected track of the target object at the current moment, the control parameters of the impedance control model and the preset double-ring impedance control model, and taking the expected joint angle as the control instruction of the double mechanical arms to realize the flexible control of the double-arm robot.
9. An electronic device comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the method of any of claims 1-7 when the computer program is executed.
10. A computer readable storage medium having stored thereon a computer program which, when executed in a computer, causes the computer to perform the method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311461575.9A CN117182929B (en) | 2023-11-06 | 2023-11-06 | Flexible control method and device for on-orbit assembly of double-arm robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311461575.9A CN117182929B (en) | 2023-11-06 | 2023-11-06 | Flexible control method and device for on-orbit assembly of double-arm robot |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117182929A true CN117182929A (en) | 2023-12-08 |
| CN117182929B CN117182929B (en) | 2024-01-23 |
Family
ID=88989054
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311461575.9A Active CN117182929B (en) | 2023-11-06 | 2023-11-06 | Flexible control method and device for on-orbit assembly of double-arm robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117182929B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170106542A1 (en) * | 2015-10-16 | 2017-04-20 | Amit Wolf | Robot and method of controlling thereof |
| CN106695797A (en) * | 2017-02-22 | 2017-05-24 | 哈尔滨工业大学深圳研究生院 | Compliance control method and system based on collaborative operation of double-arm robot |
| CN108393888A (en) * | 2017-02-06 | 2018-08-14 | 精工爱普生株式会社 | control device, robot and robot system |
| CN108638052A (en) * | 2018-03-29 | 2018-10-12 | 南京航空航天大学 | A kind of closed chain formula multi-arm robot Shared control method |
| CN110421547A (en) * | 2019-07-12 | 2019-11-08 | 中南大学 | A kind of tow-armed robot collaboration impedance adjustment based on estimated driving force model |
| CN111268182A (en) * | 2020-02-11 | 2020-06-12 | 西北工业大学 | Constraint compliance stability control method for space double-arm robot |
| CN112372637A (en) * | 2020-10-27 | 2021-02-19 | 东方红卫星移动通信有限公司 | Adaptive impedance compliance control method, module and system for low-orbit satellite space manipulator |
| CN116638507A (en) * | 2023-05-12 | 2023-08-25 | 北京理工大学 | Teleoperation control method combining self-adaptive impedance control and predictive control |
| WO2023180585A1 (en) * | 2022-03-25 | 2023-09-28 | Deepmind Technologies Limited | Controlling robots using latent action vector conditioned controller neural networks |
| CN116901057A (en) * | 2023-06-09 | 2023-10-20 | 北京理工大学 | Double mechanical arm tail end position cooperative control method based on variable impedance strategy |
-
2023
- 2023-11-06 CN CN202311461575.9A patent/CN117182929B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170106542A1 (en) * | 2015-10-16 | 2017-04-20 | Amit Wolf | Robot and method of controlling thereof |
| CN108393888A (en) * | 2017-02-06 | 2018-08-14 | 精工爱普生株式会社 | control device, robot and robot system |
| CN106695797A (en) * | 2017-02-22 | 2017-05-24 | 哈尔滨工业大学深圳研究生院 | Compliance control method and system based on collaborative operation of double-arm robot |
| CN108638052A (en) * | 2018-03-29 | 2018-10-12 | 南京航空航天大学 | A kind of closed chain formula multi-arm robot Shared control method |
| CN110421547A (en) * | 2019-07-12 | 2019-11-08 | 中南大学 | A kind of tow-armed robot collaboration impedance adjustment based on estimated driving force model |
| CN111268182A (en) * | 2020-02-11 | 2020-06-12 | 西北工业大学 | Constraint compliance stability control method for space double-arm robot |
| CN112372637A (en) * | 2020-10-27 | 2021-02-19 | 东方红卫星移动通信有限公司 | Adaptive impedance compliance control method, module and system for low-orbit satellite space manipulator |
| WO2023180585A1 (en) * | 2022-03-25 | 2023-09-28 | Deepmind Technologies Limited | Controlling robots using latent action vector conditioned controller neural networks |
| CN116638507A (en) * | 2023-05-12 | 2023-08-25 | 北京理工大学 | Teleoperation control method combining self-adaptive impedance control and predictive control |
| CN116901057A (en) * | 2023-06-09 | 2023-10-20 | 北京理工大学 | Double mechanical arm tail end position cooperative control method based on variable impedance strategy |
Non-Patent Citations (2)
| Title |
|---|
| 徐意钧等: "基于强化学习的机械臂关节高精度控制方法研究", 空间控制技术与应用, vol. 46, no. 01, pages 37 - 42 * |
| 陶东等: "模型不确定的空间机器人无力传感器阻抗控制方法", 宇航学报, vol. 42, no. 6, pages 766 - 774 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117182929B (en) | 2024-01-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107490965B (en) | Multi-constraint trajectory planning method for space free floating mechanical arm | |
| US8924021B2 (en) | Control of robots from human motion descriptors | |
| CN107505947B (en) | Space robot captured target despinning and coordination control method | |
| Flacco et al. | Prioritized multi-task motion control of redundant robots under hard joint constraints | |
| CN114952868B (en) | 7-degree-of-freedom SRS (sounding reference Signal) type mechanical arm control method and device and piano playing robot | |
| Ibanez et al. | Unified preview control for humanoid postural stability and upper-limb interaction adaptation | |
| Fan et al. | Real-time grasp planning for multi-fingered hands by finger splitting | |
| CN115781685A (en) | High-precision mechanical arm control method and system based on reinforcement learning | |
| CN112809666A (en) | 5-DOF mechanical arm force and position tracking algorithm based on neural network | |
| Zhong et al. | Toward safe human–robot interaction: A fast-response admittance control method for series elastic actuator | |
| CN113219825B (en) | Single-leg track tracking control method and system for four-leg robot | |
| CN111890363A (en) | Mechanical arm motion planning method based on rapid self-adaptive gradient neural network algorithm | |
| CN117182929B (en) | Flexible control method and device for on-orbit assembly of double-arm robot | |
| Anand et al. | Evaluation of variable impedance-and hybrid force/motioncontrollers for learning force tracking skills | |
| KR101297607B1 (en) | Method and system for robust adaptive control of multiple robots for moving target tracking and capturing | |
| JP7375587B2 (en) | Trajectory generation device, multi-link system, and trajectory generation method | |
| Pham et al. | A nonlinear model-based control of flexible robots | |
| CN113352328A (en) | Method for identifying hinge model and robot operation method | |
| CN115890735B (en) | Mechanical arm system, mechanical arm, control method of mechanical arm system, controller and storage medium | |
| Peters et al. | Learning Operational Space Control. | |
| Iqbal et al. | Arm-manipulator coordination for load sharing using predictive control | |
| Zhang et al. | Trajectory planning of collaborative robot for 3c products assembly | |
| Xue et al. | A Robot Human-Like Learning Framework Applied to Unknown Environment Interaction | |
| CN117001660A (en) | Parameter learning and synchronous control method for redundancy mechanical arm | |
| Feng et al. | Reinforcement Learning-Based Impedance Learning for Robot Admittance Control in Industrial Assembly |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |