A method for controlling flexible joints, computer equipment, and a flexible robotic arm
By establishing a relationship model between the configuration space attitude coordinates and control vectors of a flexible joint, the complexity of flexible joint control and the problem of coordinate singularities are solved, realizing a simple and universal flexible joint control method that is applicable to different hardware configurations and external disturbances.
Patent Information
- Application Number
- CN202211296243.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-10-21
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2042-10-21
AI Technical Summary
Existing flexible joint control methods are complex, lack versatility, are difficult to adapt to different hardware configurations, and are prone to errors and jitter at singularities in configuration space coordinates.
A model relating the configuration space attitude coordinates and control vectors of a flexible joint is established. The target configuration space attitude coordinates are converted into target control vectors to determine the virtual control quantity. Control commands are generated using feedback and feedforward control components to avoid the coordinate singularity problem.
It achieves simple and universal control of flexible joints, enabling stable operation across the entire motion space, thus reducing design difficulty and cost.
Smart Images

Figure CN117917309B_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of robot control technology, and in particular relates to a control method for a flexible joint, a computer device, and a flexible robotic arm. Background Technology
[0002] Flexible robotic arms are a new type of robotic arm shaped like an elephant's trunk. They are rapidly developing due to their high degree of freedom, safety, and flexibility. A flexible robotic arm is composed of multiple flexible joints connected in series; therefore, controlling and using a flexible robotic arm is based on controlling each individual flexible joint.
[0003] Currently, there are two main methods for controlling flexible joints. One method is through model-free, data-driven control. However, this method heavily relies on the hardware infrastructure of the flexible joint. For each flexible joint with a different number of fluid actuators, actuator placement, or other hardware adjustments, extensive data collection is required, making it difficult to apply in industrial settings. The other method involves establishing a kinematic model relationship between the fluid actuator space and the configuration space, and then controlling the flexible joint based on this relationship. However, this method requires complex mathematical models, analysis of flexible joints with different numbers of fluid actuators, and control of air pressure change rates. In practice, this places high demands on the equipment and is very expensive.
[0004] Besides the difficulties in control methods, the control of flexible joints also encounters the problem of handling singularities in configuration space coordinates. In configuration space, when the flexible joint is in a straight, unbent state, the bending angle in the coordinate description of the flexible joint is 0, and the bending direction is meaningless. Near the point where the bending angle is 0, even a slight change in the flexible joint's posture can cause a significant change in the coordinates of its bending direction. Therefore, the straight state of the flexible joint is the singularity position commonly described by coordinates. In closed-loop control, the singularity position can cause significant errors and jitter. Therefore, general closed-loop control avoids this point or switches to another control method in the vicinity of the singularity. However, switching control algorithms makes the overall control algorithm more complex because it is necessary to maintain smooth and continuous motion control.
[0005] It is evident that current methods for controlling flexible joints are quite complex and lack versatility. Summary of the Invention
[0006] In view of this, embodiments of this application provide a control method for a flexible joint, a computer device, and a flexible robotic arm, which can avoid the coordinate singularity problem when controlling a flexible joint, making the control of the flexible joint simpler and the flexible joint control method more versatile.
[0007] The first aspect of this application provides a method for controlling a flexible joint, including:
[0008] Establish a model relating the configuration space attitude coordinates and control vectors of a flexible joint;
[0009] Based on the target configuration space attitude coordinates of the flexible joint and the relationship model, a target control vector for the flexible joint is determined, wherein the target configuration space attitude coordinates are used to characterize the target attitude of the flexible joint.
[0010] Based on the current configuration space attitude coordinates of the flexible joint and the relationship model, the virtual control quantity of the target control vector is determined. The virtual control quantity is the input quantity for controlling the fluid actuator that makes the error value between the current control vector corresponding to the current configuration space attitude coordinates and the target control vector approach 0.
[0011] Based on the virtual control quantity and the preset control model, the control command for the flexible joint is determined;
[0012] According to the control command, the flexible joint is controlled to reach the target posture.
[0013] A second aspect of this application provides a control device for a flexible joint, comprising:
[0014] The model building module is used to establish a model of the relationship between the configuration space attitude coordinates and control vectors of the flexible joint;
[0015] The target control vector determination module is used to determine the target control vector for the flexible joint based on the target configuration space attitude coordinates of the flexible joint and the relationship model, wherein the target configuration space attitude coordinates are used to characterize the target attitude of the flexible joint.
[0016] The virtual control quantity determination module is used to determine the virtual control quantity of the target control vector based on the current configuration space attitude coordinates of the flexible joint and the relationship model. The virtual control quantity is the input quantity for controlling the fluid actuator that makes the error value between the current control vector corresponding to the current configuration space attitude coordinates and the target control vector approach 0.
[0017] The control command determination module is used to determine the control command for the flexible joint based on the virtual control quantity and the preset control model.
[0018] The control module is used to control the flexible joint to achieve the target posture according to the control command.
[0019] A third aspect of this application provides a flexible joint that is controlled by the method described in the first aspect above.
[0020] A fourth aspect of this application provides a flexible robotic arm, the flexible robotic arm including at least one flexible joint, the flexible robotic arm controlling the flexible joint by the method described in the first aspect above.
[0021] A fifth aspect of this application provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in the first aspect above.
[0022] A sixth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect above.
[0023] A seventh aspect of this application provides a computer program product that, when run on a computer device, causes the computer device to perform the method described in the first aspect above.
[0024] Compared with the prior art, the embodiments of this application have the following advantages:
[0025] In this embodiment, based on the relationship model between the configuration space attitude coordinates and control vectors of a flexible joint, the configuration space attitude coordinates can be converted into control vectors, thereby avoiding the singularity problem in the configuration space. When the flexible joint needs to change its posture, a target configuration space attitude coordinate can be given. Based on the target configuration space attitude coordinate and the relationship model, the corresponding target control vector can be determined. Then, the current configuration space coordinates of the flexible joint can be measured. Based on the current configuration space coordinates and the relationship model, the current control vector can be determined. Based on the current control vector and the target control vector, a virtual control quantity can be determined that makes the error between the current control vector and the target control vector approach 0. Based on the virtual control quantity and the preset control model, a control command for the flexible joint can be determined. Executing the control command can make the flexible joint reach the target posture corresponding to the target configuration space attitude coordinates. In this embodiment, a method for controlling a flexible joint is provided, which converts the configuration space attitude coordinates into corresponding control vectors, thereby avoiding the coordinate singularity problem in the configuration space. The control method in this embodiment can be used in various flexible joints without needing to determine a corresponding control model for each flexible joint separately. Therefore, the flexible joint control method in this application has universality. Attached Figure Description
[0026] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.
[0027] Figure 1 This is a flowchart illustrating the steps of a flexible joint control method provided in an embodiment of this application;
[0028] Figure 2 This is a schematic diagram of a flexible joint bending according to an embodiment of this application;
[0029] Figure 3 This is a schematic diagram of a flexible joint planar projection provided in an embodiment of this application;
[0030] Figure 4 This is a schematic diagram of a closed-loop control complex plane vector provided in an embodiment of this application;
[0031] Figure 5 This is a flowchart illustrating the steps of another flexible joint control method provided in this application embodiment;
[0032] Figure 6 This is a schematic diagram of closed-loop control of a flexible joint provided in an embodiment of this application;
[0033] Figure 7 This is a flowchart illustrating the steps of another flexible joint control method provided in this application embodiment;
[0034] Figure 8 This is a schematic diagram of closed-loop control of another flexible joint provided in an embodiment of this application;
[0035] Figure 9 This is a schematic diagram of a flexible robotic arm provided in an embodiment of this application;
[0036] Figure 10 This is a schematic diagram of a control device for a flexible joint provided in an embodiment of this application;
[0037] Figure 11 This is a schematic diagram of a computer device provided in an embodiment of this application. Detailed Implementation
[0038] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.
[0039] A flexible joint is a unit joint module of a flexible robotic arm. A flexible joint is generally composed of multiple fluid actuators mounted in a ring at equal intervals in parallel, typically with three or more fluid actuators per joint (N≥3). A fluid actuator is a flexible actuator that achieves linear motion by filling and releasing internal fluid. The fluid can be gas or liquid, for example, the fluid actuator can be a pneumatic linear actuator. By adjusting the pressure inside each fluid actuator, the length of each fluid actuator can be controlled to vary, thereby achieving rotation, bending, and even extension / retraction of the end face of a single joint within a limited range. For example, when there are three fluid actuators on a single joint, if any two fluid actuators are filled with fluid and become longer, and the third fluid actuator is removed and becomes shorter, the length difference causes the end face of the joint to bend towards the fluid actuator that has been removed and shortened. By adjusting the degree of filling and releasing internal fluid, the degree of bending of the end face can be controlled; by adjusting the number of fluid actuators filling and releasing internal fluid, the bending direction of the end face can be controlled, achieving control of the bending direction; if there is no mechanical constraint in the length direction, filling all three fluid actuators with internal fluid can achieve length extension / retraction of the entire joint.
[0040] Therefore, when controlling a flexible joint, it is necessary to determine the control command for each fluid actuator in the flexible joint, and control each fluid actuator according to the control command so that the flexible joint can achieve the target posture.
[0041] Because the configuration space attitude of a flexible joint requires inputs from multiple fluid actuators to collectively determine its orientation, it involves a functional mapping from one multivariate coordinate space to another. This functional mapping, referred to as the flexible joint model in this paper, is related to many factors, such as the number of fluid actuators, their placement, their output characteristics, their length, and many others. This makes it difficult to have a general model that can predict the configuration space output attitude of the flexible joint using pressure input commands from the fluid actuator space, or obtain the pressure input commands from the fluid actuator space using the configuration space attitude commands of the flexible joint. Furthermore, when the number of fluid actuators exceeds the parameters of the configuration space, for example, when the flexible joint has a central spine and its length cannot be freely extended or retracted as a degree of freedom, there will be an infinite number of sets of actuator space pressure values corresponding to the same flexible joint attitude coordinates. Therefore, additional constraints need to be introduced to determine the input pressure; generally, the stiffness of the flexible joint can be chosen as an auxiliary command. These different characteristics of flexible joints with and without a spine further increase the complexity of building flexible joint models.
[0042] The technical solution of this application will be described below through specific embodiments.
[0043] Reference Figure 1The diagram illustrates a step-by-step flowchart of a flexible joint control method provided in an embodiment of this application. This method, applied to a flexible joint, may specifically include the following steps:
[0044] S101. Establish the relationship model between the configuration space attitude coordinates and control vectors of the flexible joint. The rotation direction, bending angle, and trunk parameters of the flexible joint are called the configuration space. If the flexible joint has a spine, the trunk parameter is the stiffness of the flexible joint; if the flexible joint does not have a spine, the trunk parameter is the length of the flexible joint.
[0045] Figure 2 This is a schematic diagram of a flexible joint bending according to an embodiment of this application. Figure 2 As shown, the flexible joint can include a bottom end face and a top end face. A three-dimensional coordinate system O1X1Y1Z1 can be established with the center point O1 of the bottom end face as the origin coordinate system and the bottom end face as the xy plane. A three-dimensional coordinate system O2X2Y2Z2 can be established with the center point O2 of the top end face as the origin coordinate system and the top end face as the xy plane.
[0046] Figure 2 In this context, α represents the bending angle of the flexible joint, and β represents the bending direction of the flexible joint. For example... Figure 2 As shown, the center point O2 of the top end face can be projected onto the plane O1X1Y1 to obtain a projection point. This projection point and the center point O1 of the bottom end face can determine a ray. The angle between the ray and the ray O1X1 is the bending direction α. The angle between the ray determined by the projection point and the center point O1 of the bottom end face and the plane O2X2Y2 where the top end face is located is the bending angle β of the flexible joint.
[0047] The flexible joint may include multiple fluid actuators, which can change their shape by applying pressure to the fluid, thereby driving the target. The fluid can be liquid or air, for example, the fluid actuator in this embodiment can be a pneumatic linear actuator. Figure 2 The flexible joint in the device has three fluid actuators. Figure 3 This is a planar projection schematic diagram of a flexible joint provided in an embodiment of this application. Figure 3 It can be Figure 2 A schematic diagram of the planar projection of the flexible joint in the image. Figure 3 The mid-plane projection diagram includes the projections of three fluid actuators. The angle of each fluid actuator can be the angle between the line connecting the center point of the fluid actuator to the origin and the X-axis at the bottom end face. For example... Figure 3 As shown, the angle of fluid actuator 1 is θ1, the angle of fluid actuator 2 is θ2, and the angle of fluid actuator 3 is θ3. Figure 3 R in the figure is the vector magnitude of the bottom center point of each fluid actuator at the bottom end face of the flexible joint.
[0048] Since coordinate singularities exist in configuration space, this embodiment introduces a control vector corresponding to the configuration space attitude coordinates to avoid errors caused by these singularities. To convert configuration space attitude coordinates into control vectors, a relationship model between the configuration space attitude coordinates and the control vectors can be established.
[0049] The relational model can be:
[0050] φ=H(α,β,L)
[0051]
[0052]
[0053] Where φ is the control vector, H is the relational model, α is the bending angle of the flexible joint, β is the bending direction of the flexible joint, L is the trunk parameter of the flexible joint, and the configuration space attitude coordinates are [α,β,L]; φ c ,φ s and φ p These are the three coordinate components of the control vector, C1 and C2 are constants, k is the spring coefficient of the fluid actuator of the flexible joint, N is the number of fluid actuators in the flexible joint, A is the equivalent cross-sectional area of the fluid actuator, and R is the vector magnitude of the bottom center point of the fluid actuator at the bottom end face of the flexible joint.
[0054] S102, based on the target configuration space attitude coordinates of the flexible joint and the relationship model, determine the target control vector for the flexible joint, wherein the target configuration space attitude coordinates are used to characterize the target attitude of the flexible joint.
[0055] The aforementioned target configuration space attitude coordinates are used to characterize the desired attitude of the flexible joint after control. Inputting the target configuration space attitude coordinates into the relational model yields the corresponding target control vector, which is equivalent to transforming the target configuration space attitude coordinates into a vector space.
[0056] The target configuration space attitude coordinates can include the target bending angle, target bending direction, and target trunk parameters of the flexible joint. If the flexible joint has a spine, the target trunk parameter can be the stiffness of the flexible joint; if the flexible joint does not have a spine, the target trunk parameter can be the length of the flexible joint.
[0057] The target configuration space attitude coordinates are [α] d ,β d ,L d ], where α d For the target bending angle, β d For the target bending direction, L dThe target torso parameters are defined. The target configuration space attitude coordinates are input into the relational model to obtain the target control vector φ. d :
[0058]
[0059] The above φ cd φ sd and φ pd These are the three coordinate components of the target control vector.
[0060] S103, based on the current configuration space attitude coordinates of the flexible joint and the relationship model, determine the virtual control quantity of the target control vector. The virtual control quantity is used to make the error value between the current control vector corresponding to the current configuration space attitude coordinates and the target control vector approach 0.
[0061] When controlling a flexible joint to reach a target posture, the virtual control quantity required for the flexible joint to reach the target posture can be determined based on the current posture of the flexible joint. The virtual control quantity is the control input quantity that makes the error value between the current control vector corresponding to the current configuration space posture coordinates approach 0.
[0062] When determining the virtual control quantity, the current control vector of the flexible joint can be determined based on the current configuration space attitude coordinates of the flexible joint and the above relationship model; based on the target control vector and the current control vector, the virtual control quantity required for the flexible joint to achieve the target attitude can be determined.
[0063] In this embodiment, the control of the flexible joint is a closed-loop control method, which requires control based on negative feedback terms. These negative feedback terms can be characterized using feedback control components. The feedback control components can be determined based on the error between the current configuration space attitude and the target configuration space attitude. Based on the relational model in this embodiment, both the current configuration space attitude and the target configuration space attitude can be converted into a current control vector and a target control vector in vector space. Based on the current control vector and the target control vector, an error control vector can be determined, and then, based on the error control vector, the corresponding feedback control components can be determined.
[0064] When determining the negative feedback control components, the current control vector must first be determined. When controlling a flexible joint, it is necessary to transform the joint from its current attitude to the target attitude; therefore, the current configuration space attitude coordinates of the flexible joint can be measured. After determining the current configuration space attitude coordinates of the flexible joint, these coordinates can be input into the relational model to obtain the current control vector.
[0065] The current configuration space attitude coordinates of the flexible joint can be [α] f ,β f,L f ], where α f For the current bending angle, β f L represents the current bending direction. f The current torso parameters are defined. The current configuration space attitude coordinates are input into the relational model to obtain the current control vector φ. f It can be:
[0066]
[0067] The above φ cf φ sf and φ pf These are the three coordinate components of the current control vector.
[0068] It should be noted that when the flexible joint is of the spinal type, the current trunk parameters are generally difficult to measure. In this case, the target trunk parameters can be used as the current trunk parameters, i.e., L. f =L d The corresponding current control vector φ f It can be:
[0069] After determining the current control vector, the difference between the target control vector and the current control vector can be used as the error control vector. Specifically, the error control vector φ can be calculated based on the current control vector and the target control vector using the following formula. e for:
[0070]
[0071]
[0072]
[0073] in, ce φ se and φ pe These are the three coordinate components of the error control vector.
[0074] Figure 4 This is a schematic diagram of a complex plane vector provided in an embodiment of this application. For example... Figure 4 As shown, β f The current bending direction is represented by φ, and the corresponding current control vector is φ. f ,β d The target bending direction is φ, and the target control vector corresponding to the target bending direction is φ. d , φ e This is the error control vector. Figure 4As can be seen in the plane shown, to change the flexible joint from the current bending direction to the target bending direction, the target control vector φ can be used. d Subtract the current control vector φ f The obtained error control vector φ e To determine the necessary transformations.
[0075] When determining the feedback control components based on the error vector, the error control vector φ can be... e PID (Proportional Integral Derivative) control is performed to calculate the feedback control component U. fb for:
[0076]
[0077] in, P K i K d These are the control parameters for each sub-item, and they form a diagonal matrix.
[0078] The feedback control component obtained from the above steps can be directly used as the virtual control quantity, i.e., the virtual control quantity U = U fb .
[0079] In closed-loop control, the effect can be improved by adding a feedforward term. Therefore, in another possible implementation, the virtual control quantity can be determined jointly based on feedback control and feedforward control. Feedforward control can be characterized using feedforward control components. Feedforward control controls the flexible joint based on the target posture it needs to reach; therefore, the target control vector can be used as the feedforward control component. The sum of the feedforward control component and the feedback control component is used as the virtual control quantity. That is:
[0080] Feedforward control component U ff =φ d Virtual control quantity U = U fb +U ff .
[0081] In this embodiment of the application, when controlling a flexible joint, only the feedback control component can be used to determine the virtual control quantity; alternatively, both the feedback control component and the feedforward control component can be used to determine the virtual control quantity.
[0082] S104, Based on the virtual control quantity and the preset control model, determine the control command for the flexible joint.
[0083] When controlling flexible joints, different control methods can be used to control fluid actuators. Different control methods require different control models.
[0084] When controlling a flexible joint, pressure control can be used based on pressure feedback to control the fluid actuators. In this case, the control model can be a model of virtual control quantity and pressure control quantity. When the virtual control quantity is input into this control model, a corresponding control command is obtained, which is a pressure command. In this embodiment, the pressure can be pneumatic or hydraulic; this application uses pneumatic pressure as an example only and does not limit the scope of this application. Based on the virtual control quantity, the pressure command for each fluid actuator in the flexible joint is calculated using the following formula:
[0085]
[0086] B = Θ T (Θ T Θ) -1
[0087]
[0088] Among them, P id The pressure command for the i-th fluid actuator, U is the virtual control variable, and θ is the virtual control variable. i Let be the vector angle of the bottom center point of the i-th fluid actuator on the bottom end face of the flexible joint, where i is an integer greater than or equal to 1.
[0089] The fluid actuator has inlet and outlet valves, including an inlet valve and an outlet valve. When controlling a flexible joint, pressure feedback can be eliminated by controlling the inlet and outlet valves of the fluid actuator. In this case, the control model can be a virtual control model. When a virtual control quantity is input into this model, a corresponding control command is obtained; this control command is a virtual control command.
[0090] Based on the virtual control quantity, the virtual control command for each fluid actuator of the flexible joint is calculated using the following formula:
[0091]
[0092] B = Θ T (Θ T Θ) -1
[0093]
[0094] Where V represents the control command for the flexible joint, V id U is the virtual control command for the i-th fluid actuator, U is the virtual control variable, and θ is the virtual control command. i Let be the vector angle of the bottom center point of the i-th fluid actuator on the bottom end face of the flexible joint, where i is an integer greater than or equal to 1.
[0095] It should be noted that when using pressure control based on pressure feedback to control a fluid actuator, the virtual control quantity can be the feedback control component, or the sum of the feedback control component and the feedforward control component. When the fluid actuator is controlled by controlling its inlet and outlet valves without pressure feedback, the virtual control quantity is the feedback control component.
[0096] S105, according to the control command, control the flexible joint to reach the target posture.
[0097] After determining the control command for each fluid actuator, the corresponding fluid actuator can be controlled based on the control command, thereby enabling the flexible joint to achieve the target posture.
[0098] When controlling fluid actuators based on pressure feedback, pressure control can be applied to each fluid actuator of the flexible joint to achieve the corresponding pressure P. id This allows the flexible joint to achieve the target posture.
[0099] When fluid actuators are controlled by controlling their inlet and outlet valves without pressure feedback, the inlet and outlet valve commands for each fluid actuator can be determined based on virtual control commands. Based on these commands, the fluid actuators are controlled to achieve the target posture of the flexible joint.
[0100] Based on the virtual control commands, the inlet valve command and outlet valve command for each fluid actuator can be determined using the following formula:
[0101] O=K p2 V
[0102]
[0103] Among them, O iin For the intake valve command of the i-th fluid actuator, O iout For the outlet valve command of the i-th fluid actuator, K p2 For proportional control matrix, k i represents the diagonal parameter of the i-th row of the proportional control matrix.
[0104] In this embodiment, a relationship model between configuration space attitude coordinates and control vectors is established. Based on this model, the configuration space attitude of the flexible joint can be converted into control vectors, thereby controlling the flexible joint based on the control vectors. This avoids the problem of coordinate singularities in configuration space. The control method in this embodiment can be directly applied across the entire motion space. The relationship model in this embodiment is simple, the parameters used in the model are intuitive and readily available, and the computational load is small, making it easy to use in ordinary embedded devices.
[0105] In this embodiment, the relational model can be configured based on the parameters of the fluid actuators of the flexible joint. Therefore, the method in this embodiment can easily adapt to different numbers and installation locations of fluid actuators. Whether the flexible joint has a spine or not, the relational model can be used to convert the configuration space attitude into a control vector. Therefore, the method in this embodiment can be used for flexible joints with or without a spine. It is evident that the flexible joint control method in this embodiment has greater versatility.
[0106] Reference Figure 5 The diagram illustrates a step-by-step flowchart of another flexible joint control method provided in this application embodiment. This method is applied to a flexible joint and may specifically include the following steps:
[0107] S501, establish a relationship model between the configuration space attitude coordinates and control vectors of the flexible joint.
[0108] The execution subject in this embodiment can be a flexible joint or a flexible robotic arm containing the flexible joint. The flexible joint may include multiple fluid actuators. In this embodiment, pressure control is used to control each fluid actuator of the flexible joint, thereby enabling the flexible joint to achieve the target posture.
[0109] S502, based on the target configuration space attitude coordinates of the flexible joint and the relationship model, determine the target control vector for the flexible joint, wherein the target configuration space attitude coordinates are used to characterize the target attitude of the flexible joint.
[0110] In this embodiment, S501-S502 are similar to S101-S102 in the previous embodiment and can be referenced interchangeably. They will not be described in detail here.
[0111] S503, determine the current control vector of the flexible joint based on the current configuration space attitude coordinates of the flexible joint and the relationship model.
[0112] By using the current configuration space attitude command of the flexible joint as input to the relational model, the corresponding current control vector can be obtained.
[0113] S504, based on the current control vector and the target control vector, determine a virtual control quantity, the virtual control quantity being used to make the error value between the current control vector and the target control vector approach 0.
[0114] The control method in this embodiment is a closed-loop control method, which requires control based on negative feedback terms. These negative feedback terms can be characterized using feedback control components. The feedback control components can be determined based on the error between the current configuration space attitude and the target configuration space attitude. The error between the current configuration space attitude and the target configuration space attitude can be represented by the current control vector and the target control vector. Therefore, the difference between the target control vector and the current control vector can be used as the error control vector; then, based on the error control vector, the corresponding feedback control component is determined. This feedback control component is the virtual control quantity.
[0115] In another possible implementation, feedforward control can be added to improve the control effect. Therefore, a feedforward control component can be added when determining the virtual control quantity. Specifically, the target control vector can be used as the feedforward control component of the target control vector, and the sum of the feedforward control component and the feedback control component can be used as the virtual control quantity.
[0116] In this embodiment, the feedforward control component is optional, and the user can determine whether to add it based on their needs. In this embodiment, the feedback control component can be directly used as the virtual control quantity; alternatively, the sum of the feedback control component and the feedforward control component can be used as the virtual control quantity.
[0117] S505, based on the virtual control quantity, calculate the pressure command of each fluid actuator in the flexible joint.
[0118] S506, according to the pressure command, pressure is applied to each of the fluid actuators of the flexible joint to control the flexible joint to reach the target posture.
[0119] Based on the underlying pressure control, each fluid actuator can be controlled to reach the corresponding pressure. Under different pressures, the flexible joint can achieve the target posture, with the bending direction, bending angle, length or stiffness corresponding to the target configuration spatial posture.
[0120] This embodiment is a closed-loop control method based on complex plane vectors. The calculation of each step in this embodiment has been described in the previous embodiment and will not be repeated here.
[0121] Figure 6 This is a schematic diagram of closed-loop control for a flexible joint corresponding to an embodiment. Figure 6In this context, G represents the flexible joint to be controlled, or it can be the flexible robotic arm containing the joint, and B represents a parameter determined based on the mounting angle of the fluid actuator. For example... Figure 6 As shown, the target configuration space attitude coordinates are [α] d ,β d ,L d When input into the relational model H, the corresponding target control vector φ can be obtained. d The current configuration space pose of the flexible joint [α] f ,β f ,L f The input can be fed into the relational model H to obtain the corresponding current control vector φ. f According to the target control vector φ d and the current control vector φ f The corresponding error control vector φ can be determined. e ; for φ e PID control can yield the feedback control component U. fb Based on the target control vector φ d The feedforward control component U can be obtained. ff Based on feedback control component U fb and feedforward control component U ff The virtual control quantity U is obtained. Based on the virtual control quantity U and parameter B, the pressure command P for each fluid actuator can be obtained. id According to pressure command P id The flexible joint G can be controlled.
[0122] In this embodiment, closed-loop control of the flexible joint can be performed based on both feedforward control and feedback control, resulting in good and stable control performance that can adapt to different loads and external disturbances.
[0123] Reference Figure 7 The diagram illustrates a step-by-step flowchart of another flexible joint control method provided in this application embodiment. This method is applied to a flexible joint and may specifically include the following steps:
[0124] S701, establish a relationship model between the configuration space attitude coordinates and control vectors of the flexible joint.
[0125] The execution subject in this embodiment can be a flexible joint or a flexible robotic arm in which the flexible joint is located.
[0126] Existing control methods for flexible joints all require independent pressure feedback for each fluid actuator. For a flexible continuous robotic arm containing multiple flexible joints, the number of pressure sensors required for pressure feedback would be dozens or more. This increases the difficulty of installing pressure sensors, the design complexity of flexible joints, and the cost of the flexible continuous robotic arm.
[0127] Based on this, this embodiment proposes a flexible joint control method that does not require pressure feedback.
[0128] The fluid actuator includes an inlet valve and an outlet valve. In this embodiment, the flexible joint is controlled by controlling the inlet valve and the outlet valve.
[0129] S702, Based on the target configuration space attitude coordinates of the flexible joint and the relationship model, determine the target control vector for the flexible joint, wherein the target configuration space attitude coordinates are used to characterize the target attitude of the flexible joint.
[0130] In this embodiment, S701-S702 are similar to S101-S102 in the previous embodiment and can be referenced interchangeably; therefore, they will not be described in detail here. The flexible joint may include multiple fluid actuators. In this embodiment, pressure control is used to control each fluid actuator of the flexible joint, thereby enabling the flexible joint to achieve the target posture.
[0131] S703, determine the current control vector of the flexible joint based on the current configuration space attitude coordinates of the flexible joint and the relationship model.
[0132] S704, determine the feedback control component of the target control vector based on the current control vector and the target control vector.
[0133] S705, the feedback control component is used as the virtual control quantity.
[0134] S706, based on the virtual control quantity, calculate the virtual control command for each fluid actuator of the flexible joint.
[0135] S707, based on the virtual control command, determine the inlet valve command and outlet valve command for each of the fluid actuators.
[0136] Based on the virtual control quantity, the virtual control command for each fluid actuator of the flexible joint can be calculated using the following formula:
[0137]
[0138] B = Θ T (Θ T Θ) -1
[0139]
[0140] Where V represents the control command for the flexible joint, V id U is the virtual control command for the i-th fluid actuator, U is the virtual control variable, and θ is the virtual control command. i Let be the vector angle of the bottom center point of the i-th fluid actuator on the bottom end face of the flexible joint, where i is an integer greater than or equal to 1.
[0141] S708, based on the inlet valve command and the outlet valve command, controls the operation of the fluid actuator to make the flexible joint reach the target posture.
[0142] By controlling the virtual control command V of the fluid actuator, and through the relationship model Q between flow rate and valve opening size, the opening command O of the inlet and outlet valves of each fluid actuator can be obtained. Here, Q can be used as the simplest proportional control model, and the opening command O is calculated as follows:
[0143] O=K p2 V
[0144] Where K P2 This is the proportional control matrix, which is a diagonal matrix. The proportional control matrix represents system parameters that can be adjusted.
[0145] Specifically, the intake valve command O of the i-th linear actuator iin and the air outlet valve command O iout They are respectively
[0146]
[0147] Where k i represents the diagonal parameter of the i-th row of the proportional control matrix.
[0148] This embodiment is a closed-loop control method based on feedback control components that does not require pressure feedback. The calculation of each step in this embodiment has been described in the first embodiment and will not be repeated here.
[0149] Figure 8 This is a schematic diagram of closed-loop control for a flexible joint corresponding to an embodiment. Figure 8 In this context, G represents the flexible joint to be controlled, or it can be the flexible robotic arm containing the joint, and B represents a parameter determined based on the mounting angle of the fluid actuator. For example... Figure 8 As shown, the target configuration space attitude coordinates are [α] d ,β d ,L d When input into the relational model H, the corresponding target control vector φ can be obtained. dThe current configuration space pose of the flexible joint [α] f ,β f ,L f The input can be fed into the relational model H to obtain the corresponding current control vector φ. f According to the target control vector φ d and the current control vector φ f The corresponding error control vector φ can be determined. e ; for φ e PID control can yield the feedback control component U. fb Based on feedback control component U fb With parameter B, the virtual command V for each fluid actuator can be obtained. id According to V id The model Q determines the inlet valve command and the outlet valve command. Based on the inlet valve command and the outlet valve command, each fluid actuator can be controlled so that the flexible joint G can reach the corresponding target posture.
[0150] In this embodiment, pressure feedback is not required when controlling the flexible joint. Valve control commands for the inlet and outlet of the fluid actuator can be directly generated, thereby eliminating the need for pressure sensors in the flexible joint control process and reducing the design complexity and overall cost of the flexible joint. The flexible joint control method in this embodiment is closed-loop control, which can adapt to different loads and external disturbances.
[0151] It should be noted that the fluid actuator of this application can be gas-driven, such as air, or liquid-driven, such as water or oil. No specific limitation is made here. This application only uses gas-driven as an example for illustration and does not limit this application.
[0152] The flexible joint of this application may be composed of multiple fluid actuators as described above, and the flexible robotic arm of this application may be composed of multiple flexible joints as described above.
[0153] It should be noted that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.
[0154] This application embodiment also provides a flexible joint, which can be as follows: Figure 2 As shown, the flexible joint is controlled through the steps in the above method embodiments.
[0155] Reference Figure 9 This illustration shows a flexible robotic arm provided in an embodiment of this application. The flexible robotic arm can be a flexible continuous robotic arm. For example... Figure 9As shown, the flexible robotic arm may include multiple flexible joints, which can be the flexible joints described in the foregoing embodiments. By controlling the flexible joints, the flexible robotic arm can achieve a target posture or perform a target movement. The flexible robotic arm can control each flexible joint using the steps described in the foregoing method embodiments.
[0156] Reference Figure 10 The diagram illustrates a control device for a flexible joint according to an embodiment of this application. Specifically, it may include a model building module 1001, a target control vector determination module 1002, a virtual control quantity determination module 1003, a control command determination module 1004, and a control module 1005, wherein:
[0157] Model building module 1001 is used to establish a model of the relationship between the configuration space attitude coordinates and control vectors of the flexible joint;
[0158] The target control vector determination module 1002 is used to determine the target control vector for the flexible joint based on the target configuration space attitude coordinates of the flexible joint and the relationship model, wherein the target configuration space attitude coordinates are used to characterize the target attitude of the flexible joint.
[0159] The virtual control quantity determination module 1003 is used to determine the virtual control quantity of the target control vector based on the current configuration space attitude coordinates of the flexible joint and the relationship model. The virtual control quantity is the input quantity for controlling the fluid actuator that makes the error value between the current control vector corresponding to the current configuration space attitude coordinates and the target control vector approach 0.
[0160] The control command determination module 1004 is used to determine the control command for the flexible joint based on the virtual control quantity and the preset control model.
[0161] The control module 1005 is used to control the flexible joint to achieve the target posture according to the control command.
[0162] In one possible implementation, the relationship between the configuration space attitude coordinates and control vectors of the flexible joint is modeled as follows:
[0163] φ=H(α,β,L)
[0164]
[0165]
[0166] Where φ is the control vector, H is the relational model, α is the bending angle of the flexible joint, β is the bending direction of the flexible joint, L is the trunk parameter of the flexible joint, and the configuration space attitude coordinates are [α,β,L]; φ c ,φ s and φ p These are the three coordinate components of the control vector, C1 and C2 are constants, k is the spring coefficient of the fluid actuator of the flexible joint, N is the number of fluid actuators in the flexible joint, A is the equivalent cross-sectional area of the fluid actuator, and R is the vector magnitude of the bottom center point of the fluid actuator on the bottom end face of the flexible joint.
[0167] In one possible implementation, the virtual control quantity determination module 1003 includes:
[0168] The current control vector determination submodule is used to determine the current control vector of the flexible joint based on the current configuration space attitude coordinates of the flexible joint and the relationship model.
[0169] The virtual control quantity determination submodule is used to determine the virtual control quantity required for the flexible joint to achieve the target posture based on the target control vector and the current control vector.
[0170] In one possible implementation, the aforementioned virtual control quantity determination submodule includes:
[0171] An error control vector determination unit is used to determine an error control vector based on the target control vector and the current control vector;
[0172] The feedback control component determination unit is used to determine the feedback control component of the target control vector based on the error control vector.
[0173] A virtual control quantity determination unit is used to determine the virtual control quantity based on the feedback control component.
[0174] In one possible implementation, the flexible joint includes multiple fluid actuators, and the control model is used to characterize the relationship between the virtual control quantity and the valve control quantity of the fluid actuators. The control command determination module 1004 includes:
[0175] The virtual control command calculation submodule is used to calculate the virtual control command for each fluid actuator of the flexible joint based on the virtual control quantity using the following formula:
[0176]
[0177] B = Θ T (Θ T Θ)-1
[0178]
[0179] Wherein, V represents the control command for the flexible joint, V id The virtual control command for the i-th fluid actuator, U is the virtual control variable, and θ is the virtual control variable. i Let i be the vector angle of the bottom center point of the i-th fluid actuator on the bottom end face of the flexible joint, where i is an integer greater than or equal to 1.
[0180] In one possible implementation, the fluid actuator includes an inlet valve and an outlet valve, and the control module 1005 includes:
[0181] The valve command determination submodule is used to determine the inlet valve command and outlet valve command for each of the fluid actuators based on the virtual control command.
[0182] The control submodule is used to control the operation of the fluid actuator based on the inlet valve command and the outlet valve command, so that the flexible joint reaches the target posture.
[0183] In one possible implementation, the valve command determination submodule mentioned above includes:
[0184] The valve command calculation unit is used to determine the inlet valve command and outlet valve command for each fluid actuator based on the virtual control command using the following formula:
[0185] O=K p2 V
[0186]
[0187] Among them, O iin For the intake valve command of the i-th fluid actuator, O iout For the air outlet valve command of the i-th fluid actuator, K p2 This is a proportional control matrix.
[0188] In one possible implementation, the aforementioned virtual control quantity determination submodule includes:
[0189] An error control vector determination unit is used to determine an error control vector based on the target control vector and the current control vector;
[0190] A control component determination unit is used to determine the feedback control component of the target control vector based on the error control vector, and to determine the feedforward control component of the target control vector based on the target control vector.
[0191] A virtual control quantity determination unit is used to determine the virtual control quantity based on the feedback control component and the feedforward control component.
[0192] In one possible implementation, the control model is used to characterize the relationship between the virtual control quantity and the pressure control quantity, and the control command determination module 1004 includes:
[0193] The pressure command calculation submodule is used to calculate the pressure command of each fluid actuator in the flexible joint according to the virtual control quantity using the following formula:
[0194]
[0195] B = Θ T (Θ T Θ) -1
[0196]
[0197] Among them, P id For the i-th fluid actuator, U is the virtual control variable, and θ is the pressure command. i Let i be the vector angle of the bottom center point of the i-th fluid actuator on the bottom end face of the flexible joint, where i is an integer greater than or equal to 1.
[0198] In one possible implementation, the control module 1005 includes:
[0199] The pressure control submodule is used to apply pressure to each of the fluid actuators of the flexible joint according to the pressure command, so as to control the flexible joint to reach the target posture.
[0200] In one possible implementation, the control component determination unit or the feedback control component determination unit includes:
[0201] The feedback control component calculation subunit is used to calculate the feedback control component of the target control vector based on the error control vector using the following formula:
[0202]
[0203] Among them, U fb For the feedback control component, φ e Let K be the error control vector. P K i K d These are the control parameters.
[0204] As the apparatus embodiments are basically similar to the method embodiments, they are described in a relatively simple manner. For relevant details, please refer to the description in the method embodiment section.
[0205] Figure 10 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Figure 10 As shown, the computer device 110 of this embodiment includes: at least one processor 1100 ( Figure 10 (Only one is shown) a processor, a memory 1101, and a computer program 1102 stored in the memory 1101 and executable on the at least one processor 1100, which, when executing the computer program 1102, implements the steps in any of the above method embodiments.
[0206] The computer device may include, but is not limited to, a processor 1100 and a memory 1101. Those skilled in the art will understand that... Figure 11 The computer device 110 is merely an example and does not constitute a limitation on the computer device 110. It may include more or fewer components than shown, or combine certain components, or different components, such as input / output devices, network access devices, etc.
[0207] The processor 1100 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.
[0208] In some embodiments, the memory 1101 may be an internal storage unit of the computer device 110, such as a hard disk or memory of the computer device 110. In other embodiments, the memory 1101 may be an external storage device of the computer device 110, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 110. Further, the memory 1101 may include both internal and external storage units of the computer device 110. The memory 1101 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 1101 can also be used to temporarily store data that has been output or will be output.
[0209] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.
[0210] This application provides a computer program product that, when run on a computer device, enables the computer device to perform the steps described in the above-described method embodiments.
[0211] The embodiments described above are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.
Claims
1. A method for controlling a flexible joint, characterized in that, include: A relationship model between the configuration space attitude coordinates and control vectors of the flexible joint is established. The relationship model between the configuration space attitude coordinates and control vectors of the flexible joint is as follows: , in, Let H be the control vector, and H be the relation model. The bending angle of the flexible joint. The bending direction of the flexible joint. The torso parameters of the flexible joint are given, and the configuration space attitude coordinates are given. ; , and These are the three coordinate components of the control vector. and All are constants. Let N be the spring coefficient of the fluid actuator of the flexible joint, N be the number of fluid actuators in the flexible joint, A be the equivalent cross-sectional area of the fluid actuator, and R be the vector magnitude of the bottom center point of the fluid actuator on the bottom end face of the flexible joint. Based on the target configuration space attitude coordinates of the flexible joint and the relationship model, a target control vector for the flexible joint is determined, wherein the target configuration space attitude coordinates are used to characterize the target attitude of the flexible joint. Based on the current configuration space attitude coordinates of the flexible joint and the relationship model, a virtual control quantity of the target control vector is determined. The virtual control quantity is used to make the error value between the current control vector corresponding to the current configuration space attitude coordinates and the target control vector approach 0. Based on the virtual control quantity and the preset control model, the control command for the flexible joint is determined; According to the control command, the flexible joint is controlled to reach the target posture.
2. The control method as described in claim 1, characterized in that, The step of determining the virtual control quantity of the target control vector based on the current configuration space attitude coordinates of the flexible joint and the relationship model includes: Based on the current configuration space attitude coordinates of the flexible joint and the relationship model, determine the current control vector of the flexible joint; Based on the target control vector and the current control vector, determine the virtual control amount required for the flexible joint to achieve the target posture.
3. The control method as described in claim 2, characterized in that, The step of determining the virtual control quantity required for the flexible joint to achieve the target posture based on the target control vector and the current control vector includes: Based on the target control vector and the current control vector, the error control vector is determined; Based on the error control vector, determine the feedback control component of the target control vector; The virtual control quantity is determined based on the feedback control component.
4. The control method as described in claim 3, characterized in that, The flexible joint includes multiple fluid actuators. The control model is used to characterize the relationship between the virtual control quantity and the valve control quantity of the fluid actuators. Determining control commands for the flexible joint based on the virtual control quantity and the preset control model includes: Based on the virtual control quantity, the virtual control command for each fluid actuator of the flexible joint is calculated using the following formula: in, For control commands targeting the flexible joint, For the i-th fluid actuator, The virtual control quantity, Let i be the vector angle of the bottom center point of the i-th fluid actuator on the bottom end face of the flexible joint, where i is an integer greater than or equal to 1.
5. The control method as described in claim 4, characterized in that, The fluid actuator includes an inlet valve and an outlet valve. Controlling the flexible joint to achieve the target posture according to the control command includes: Based on the virtual control commands, determine the inlet valve command and outlet valve command for each of the fluid actuators; Based on the inlet valve command and the outlet valve command, the fluid actuator is controlled to operate so that the flexible joint reaches the target posture.
6. The control method as described in claim 5, characterized in that, The step of determining the inlet valve command and outlet valve command for each fluid actuator based on the virtual control command includes: Based on the virtual control commands, the inlet valve command and outlet valve command for each fluid actuator are determined using the following formula: , in, The intake valve command for the i-th fluid actuator. The command for the outlet valve of the i-th fluid actuator. This is a proportional control matrix.
7. The control method as described in claim 2, characterized in that, The step of determining the virtual control quantity required for the flexible joint to achieve the target posture based on the target control vector and the current control vector includes: Based on the target control vector and the current control vector, the error control vector is determined; The feedback control component of the target control vector is determined based on the error control vector, and the feedforward control component of the target control vector is determined based on the target control vector. The virtual control quantity is determined based on the feedback control component and the feedforward control component.
8. The control method as described in claim 3 or 7, characterized in that, The control model is used to characterize the relationship between the virtual control quantity and the pressure control quantity. The flexible joint includes multiple fluid actuators. Determining the control command for the flexible joint based on the virtual control quantity and the preset control model includes: Based on the virtual control quantity, the pressure command for each fluid actuator in the flexible joint is calculated using the following formula: in, For the pressure command of the i-th fluid actuator, The virtual control quantity, Let i be the vector angle of the bottom center point of the i-th fluid actuator on the bottom end face of the flexible joint, where i is an integer greater than or equal to 1.
9. The control method as described in claim 8, characterized in that, The step of controlling the flexible joint to reach the target posture according to the control command includes: According to the pressure command, each of the fluid actuators of the flexible joint is pressure-driven to control the flexible joint to reach the target posture.
10. The control method as described in claim 3 or 7, characterized in that, The step of calculating the feedback control component of the target control vector based on the error control vector includes: Based on the error control vector, the feedback control component of the target control vector is calculated using the following formula: in, The feedback control component, Let be the error control vector. , , These are the control parameters.
11. A flexible joint, characterized in that, The flexible joint is controlled by the method described in any one of claims 1-10.
12. A flexible robotic arm, characterized in that, The flexible robotic arm includes at least one flexible joint, and the flexible robotic arm controls the flexible joint by means of the method described in any one of claims 1-10.
13. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the control method as described in any one of claims 1-10.
14. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the control method as described in any one of claims 1-10.
Citation Information
Patent Citations
Variable-rigidity flexible mechanical arm module and control method thereof
CN109648550A
Continuous robot arm type self-estimation method
CN111618824A
