Force-position mixed five-finger dexterous hand feeling control method and system

Abstract

The invention discloses a force-position mixed five-finger dexterous hand feeling control method and system, and the method comprises the steps: employing a modular design scheme for magnetic coding sensors, and enabling the magnetic coding sensors of three joints on each finger to be integrated on a flexible circuit board; two strain gauges are respectively mounted at the fingertip joints of the fingers in the horizontal direction and the vertical direction to form two stress measurement bridges; joint position information and fingertip moment information are obtained through magnetic coding sensors on finger joints and strain gauges on fingertips, and the method is used for data output of the force-position hybrid control method for the fingers of the dexterous hand. The implementation cost of the dexterous hand is far lower than that of other schemes, the manufacturing cost of the dexterous hand is reduced, mass production is facilitated, the structure is simpler, the assembly difficulty is lower, and mass production is also facilitated.

Description

Technical Field

[0001] This invention relates to the field of dexterity technology, specifically to a force-position hybrid five-finger dexterity hand sensing and control method and system. Background Technology

[0002] The human hand has over 20 degrees of freedom, enabling it to perform fine and complex movements such as grasping, pressing, and sliding. Achieving similar functionality in a robotic system requires developing a dexterous hand hardware system with multiple degrees of freedom. Such a system not only needs a flexible mechanical structure design but also requires sensors to perceive the movement of the fingers and their contact with the external environment in real time. The challenges of a five-finger dexterous hand system include how to precisely control the movement of each finger joint and how to ensure the robot can dynamically adjust based on sensory feedback during task execution. A dexterous hand needs to be equipped with force sensors to sense the force when the fingers contact objects, position sensors to sense the angle and position of each finger joint, and a sophisticated control algorithm to effectively fuse joint position information with fingertip force information, outputting a precise motor control value to ensure high positional accuracy and force compliance in finger movements.

[0003] Traditional dexterous hand drive systems, constrained by size limitations, often only consider the motor drive system, with limited design for the dexterous hand's sensory system. They only know their joint angles and calculated fingertip positions, but have no idea whether the fingertips have touched the object or the amount of force applied. This results in either abrupt stopping before contact (inoperability) or crashing into the object with excessive force (unsafe). Furthermore, current dexterous hand drive systems suffer from the following shortcomings: 1. Limitations of designing only one sensor per finger The fingers have insufficient ability to perceive bending angles, making them unable to cope with complex application environments. They require complex angle calculation algorithms and cannot achieve precise control of the finger bending angle.

[0004] 2. Sensor usage Stress-free sensors: cannot sense force, and cannot form an effective closed loop for the gripping force of objects.

[0005] Torque sensors: expensive, require large structural space, and waste valuable internal space of dexterous hands.

[0006] Sensor arrays are expensive, difficult to deploy, have complex signal processing, and are susceptible to external interference.

[0007] 3. Each sensor has its own processing circuit and its own cable.

[0008] Due to the unique structural characteristics of dexterous hands, each sensor has a different external size, resulting in a wide variety of internal circuit boards and high maintenance costs. Furthermore, traditional cable solutions have numerous connection points (sensor solder joints, cable connectors), each a potential source of failure (poor soldering, poor contact, cable fatigue). Since fingers are continuously moving parts, traditional cables are prone to internal wire breakage under repeated bending.

[0009] 4. Simple position control Because its control loop only cares about the positional error, it continuously outputs torque to "correct" the positional deviation caused by the contact, resulting in a rigid collision that is destructive. Summary of the Invention

[0010] The technical problem to be solved by this invention is to provide a low-cost, highly compliant finger control and a force-position hybrid five-finger dexterity hand control method with strong sensing ability.

[0011] To solve the above-mentioned technical problems, the present invention provides the following technical solution: A force-position hybrid five-finger dexterity tactile control method includes: The magnetic coding sensor adopts a modular design scheme, integrating the magnetic coding sensors of three joints on each finger onto a flexible circuit board; Two strain gauges are attached to the fingertip joint in the horizontal direction and two in the vertical direction to form a two-way stress measurement bridge. The joint position information and fingertip torque information are obtained by magnetic coding sensors on the finger joints and strain gauges on the fingertips, which are used for the data output of the force-position hybrid control method for dexterous hand fingers.

[0012] In this embodiment, the flexible circuit board includes a slender main circuit flexible board 10, a first slender branch circuit flexible board 11, and a second slender branch circuit flexible board 12; the magnetic coding sensors set from the fingertip joint and the two joints below the fingertip joint are defined as the first magnetic coding sensor 21, the second magnetic coding sensor 22, and the third magnetic coding sensor 23 in sequence. Among them, the slender main circuit flexible board 10 is attached along the surface of the bionic finger, and one end of the slender main circuit flexible board 10 is defined as the main end 101 and the other end as the main root 102; the first magnetic coding sensor 21 is disposed on the main end 101 and attached to the fingertip joint of the bionic finger, and the main root 102 of the slender main circuit flexible board 10 is provided with a gold finger interface 13. The first slender branch circuit flexible plate 11 and the second slender branch circuit flexible plate 12 are arranged along both sides of the slender main circuit flexible plate 10; and the second magnetic coding sensor 22 and the third magnetic coding sensor 23 are respectively arranged at the ends of the first slender branch circuit flexible plate 11 and the second slender branch circuit flexible plate 12, and respectively attached to the two remaining joints of the bionic finger except for the fingertip joint. Electrical wiring between the three magnetically encoded sensors and the gold finger interface 13 is integrated on a flexible circuit board.

[0013] In this embodiment, the flexible circuit board is locally reinforced at the fixed position of the magnetic coding sensor to increase the thickness of the flexible circuit board.

[0014] In this embodiment, the main stem end 101 is stepped.

[0015] In this embodiment, the strain gauge is installed as a separate module at the fingertip joint.

[0016] In this embodiment, the data output of the force-position hybrid control method for dexterous hand fingers includes: The three joints of the fingers of a dexterous hand are simplified into a three-bar linkage model; The current fingertip position is obtained based on the angle information of the magnetically encoded sensors on the three joints and the equivalent link length; The current fingertip speed is calculated by numerically differencing historical values ​​of fingertip position information. Define a virtual spring and damper, and obtain control correction values ​​based on the current fingertip speed and fingertip torque information; Based on the control correction amount for the original fingertip Perform offset correction to obtain a new, compliant target position; The target values ​​of joint angles for each joint are obtained by inverse kinematics calculation of the new, compliant target position. Based on the target joint angle values ​​of each joint, a PID controller is used to control the joint motor actuator.

[0017] In this embodiment, obtaining the current fingertip position includes: Model the motion position and posture of the bending degree of freedom of a single finger in a dexterous hand. Establish a single finger joint coordinate system for the structure of a single finger and obtain the finger structure parameters of each joint segment. Substitute the finger structure parameters of each joint segment into the DH coordinate system and change them to obtain the pose change matrix of each phalanx. According to the chain rule of coordinate system transformation, the pose transformation matrix from the fingertip coordinate system to the finger root base coordinate system can be obtained. The current position of the fingertip is obtained based on the pose transformation matrix from the fingertip coordinate system to the finger root base coordinate system.

[0018] In this embodiment, the current position of the fingertip is represented by the following formula: ; ; In the formula, The position of the front fingertip, , , These are the lengths of the three joint segments from the fingertip to the base of the finger. , , These represent the rotation angles of the three joint segments from the fingertip to the base of the finger around the Z-axis, where cos is the cosine function and sin is the sine function.

[0019] In this embodiment, the new, compliant target position is obtained using the following formula: ; ; In the formula, For the new, compliant target location, For the original fingertips , To control the amount of correction, For virtual stiffness, For fingertip torque information, For virtual damping, This represents the current fingertip speed.

[0020] The present invention also provides a system applying the above-described force-position hybrid five-finger dexterity hand-feel control method, comprising: The MCU main controller is used to acquire joint position information and fingertip torque information, and as a calculation unit, it acquires data output for the force-position hybrid control method. The joint motor actuator is connected to the MCU master controller and is used to execute the data output by the force-position hybrid control method according to the control of the MCU master controller.

[0021] Compared with the prior art, the beneficial effects of the present invention are: 1. Solved the problem of insufficient sensory ability in dexterous hands. Traditional dexterous hand drive systems, constrained by size, often only consider the motor drive system, with little design for the dexterous hand's sensing system. It only knows its joint angles and the calculated fingertip position, but has no idea whether the fingertip has touched the object or with what force. This leads to either abrupt stopping before contact (inoperability) or impacting the object with excessive force (unsafe). By reading the force signal from strain gauges in real time, the control system can immediately detect unexpected increases in contact force. Combined with the current position provided by a magnetically coded sensor, the controller can instantly switch from a high-stiffness position control mode to a low-stiffness impedance control mode or directly release the force. The hand will behave "compliantly," yielding or rebounding upon contact, limiting the impact force to a safe range.

[0022] 2. The need for highly flexible finger control Traditional rigid position control generates enormous impact force in the event of accidental contact. Force-position hybrid compliant control, by drastically reducing its stiffness upon detecting or anticipating contact, allows the finger to "yield" or "rebound," limiting the impact force to a safe range and transforming the collision problem into a compliant contact problem.

[0023] 3. Maintenance and development costs The sensor solution using magnetically encoded sensors and strain gauges in dexterous hands offers a significant cost advantage compared to other more advanced solutions (such as integrating torque sensors into each joint or using six-dimensional force sensors in the fingertips), while simultaneously achieving a very high level of control performance. Through ingenious system design and algorithms, it achieves force sensing capabilities approaching those of high-end solutions using a relatively low-cost sensor combination.

[0024] Strain gauge sensors are typically mounted as a separate module on the fingertip. If damaged, this module can be replaced individually without disassembling the entire complex joint actuator. In contrast, torque sensors integrated inside the joint often require replacement of the entire joint module if damaged, resulting in significantly higher repair costs and downtime.

[0025] The solution of integrating three joint magnetic encoder sensors onto a single FPC circuit board encapsulates all electrical traces on a thin, flexible film, typically less than 0.2mm thick. This significantly saves valuable space within the finger, freeing up room for other components such as motors, reducers, and bearings. The entire sensor module (FPC + encoder) can be manufactured and tested as a standalone sub-module. During finger assembly, simply inserting one FPC module into the corresponding gold finger socket completes the electrical connection of all encoders. This "plug-and-play" approach greatly simplifies the assembly process, reduces labor costs, and is ideal for automated production. More importantly, the FPC can be customized to fit the finger's internal geometry, perfectly conforming to the finger structure, bypassing obstacles, and achieving optimal space utilization. This is something that rigid PCBs or discrete wiring cannot achieve.

[0026] This invention provides a highly integrated encoder + strain gauge solution, integrating 15-channel magnetic encoder signal acquisition and 10-channel strain gauge signal conditioning circuitry onto a miniaturized PCB. Through multi-layer layout and precise wiring design, the overall size is effectively reduced and anti-interference capability is improved. By combining "centralized" precise force sensing (fingertip) and "distributed" position sensing (each joint), along with powerful real-time algorithms (dynamics and kinematics), a force control system capable of handling most compliant grasping tasks is constructed, while cleverly avoiding the need to deploy high-cost torque sensors at every joint.

[0027] The strain gauge solution of the present invention has two major advantages over existing solutions that use multi-dimensional force-tactile sensors (such as three-dimensional force-tactile sensors): 1. The cost is around ten yuan, while three-dimensional force tactile sensors generally cost more than a thousand yuan. Each dexterous hand needs five sensors. The cost advantage of this invention is obvious and it is more suitable for mass production.

[0028] 2. Advantages in size: Three-dimensional force structures are complex and bulky, making them unsuitable for mounting between dexterous hands or inside fingers. Strain gauges, on the other hand, have a flexible structure and can be easily attached to the crossbeams inside finger joints.

[0029] The compliant control scheme for the dexterous hand in this invention requires only an encoder and a stress-strain sensor. Current control schemes often integrate multi-source data collected by sensor arrays to achieve high precision; however, this approach results in too many types and a large number of sensors, leading to complex installation, hindering mass production, and significantly increasing the computational load on the processor, thus reducing system robustness. The control scheme proposed in this invention has lower implementation complexity and consumes less computing power, making it highly advantageous in the deployment of embodied intelligence-based whole-machine control computing power.

[0030] In summary, the advantages of this invention lie in its significantly lower implementation cost compared to other solutions, reducing the manufacturing cost of the dexterous hand and facilitating mass production. Its simpler structure and easier assembly also contribute to mass production. Compared to compliant control algorithms based on multi-sensor fusion, the compliant control algorithm based on stress-strain sensors and magnetically coded sensors requires less computing power, has lower system complexity, and improves the maintainability and reliability of the entire device. Attached Figure Description

[0031] Figure 1 This is a flowchart of a force-position hybrid five-finger dexterity hand sensing control method according to an embodiment of the present invention.

[0032] Figure 2 This is a schematic diagram of the magnetic coding sensor and flexible circuit board according to an embodiment of the present invention.

[0033] Figure 3 This is a schematic diagram of strain gauge installation according to an embodiment of the present invention.

[0034] Figure 4 This is a schematic diagram of the stress measurement circuit according to an embodiment of the present invention.

[0035] Figure 5 This is a flowchart of sensor data processing according to an embodiment of the present invention.

[0036] Figure 6 This is a schematic diagram of the DH coordinate system of the bionic finger in an embodiment of the present invention.

[0037] Figure 7 This is a flowchart illustrating the data output of the force-position hybrid control method for the fingers of a dexterous hand according to an embodiment of the present invention.

[0038] Figure 8 This is a block diagram of a force-position hybrid five-finger dexterity hand control system according to an embodiment of the present invention. Detailed Implementation

[0039] To facilitate understanding of the technical solution of the present invention by those skilled in the art, the technical solution of the present invention will now be further described in conjunction with the accompanying drawings.

[0040] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0041] Please see Figure 1 As shown, the present invention provides a force-position hybrid five-finger dexterity tactile control method, comprising: The S10 magnetic coding sensor adopts a modular design, integrating the magnetic coding sensors of the three joints of each finger onto a flexible circuit board.

[0042] Please see Figure 2 As shown, in one embodiment of the present invention, the magnetic encoder sensor adopts a modular design, integrating the encoders for the three joints of each finger onto a flexible circuit board. In this embodiment, the flexible circuit board refers to an FPC (Flexible Printed Circuit) circuit board.

[0043] In this embodiment, the flexible circuit board includes a slender main circuit flexible board 10, a first slender branch circuit flexible board 11, and a second slender branch circuit flexible board 12. Magnetic coding sensors positioned from the fingertip joint and the two joints downwards from the fingertip joint are defined as a first magnetic coding sensor 21, a second magnetic coding sensor 22, and a third magnetic coding sensor 23, respectively.

[0044] The slender, flexible main circuit board 10 is attached to the surface of the bionic finger. One end of the slender, flexible main circuit board 10 is defined as the main end 101, and the other end is defined as the main root 101. A first magnetically encoded sensor 21 is disposed on the main end 101 and attached to the fingertip joint of the bionic finger. A gold finger interface 13 is provided on the main root 102 of the slender, flexible main circuit board 10. Furthermore, the main end 101 is stepped to increase the sensor area and ensure installation strength. The connection part is designed with a minimum width to facilitate wiring inside the finger.

[0045] The first and second slender branch circuit flexible boards 11 and 12 are arranged along both sides of the slender main circuit flexible board 10, making it less likely for the two slender branch circuit flexible boards to collide, thus improving safety. The second magnetic encoding sensor 22 and the third magnetic encoding sensor 23 are respectively disposed at the ends of the first and second slender branch circuit flexible boards 11 and 12, and are respectively attached to the two remaining joints of the bionic finger, excluding the fingertip joint. The electrical wiring between the three magnetic encoding sensors and the gold finger interface 13 is integrated on the flexible circuit board.

[0046] In this embodiment, the flexible circuit board (FPC) is locally reinforced at the fixed position of the magnetic encoder sensor, increasing the thickness of the FPC. This local reinforcement ensures sufficient strength at the mounting location of the magnetic encoder sensor, while the FPC itself possesses excellent flexibility, allowing it to fit closely to the internal structure of the finger and minimize the impact during finger movement. Furthermore, the five fingers maintain consistency, have a unified interface, are interchangeable, and offer high maintainability.

[0047] S20: Two strain gauges are attached to the fingertip joint in both the horizontal and vertical directions to form a two-way stress measurement bridge.

[0048] In one embodiment of the present invention, Figure 3 A crossbeam for finger joints is assembled at the fingertip joint. Two strain gauges 30 are attached horizontally and vertically, forming a two-channel stress measurement bridge. By analyzing the measurement data from the strain gauges 30, the forces acting on the fingertip in both directions can be calculated. The stress measurement bridge refers to connecting the strain gauges to a Wheatstone bridge circuit to convert minute resistance changes (corresponding to strain / stress) into voltage signals. This application does not limit the specific structure of the stress measurement bridge. Please refer to [link to relevant documentation]. Figure 4 The finger stress measurement circuit features a dual-channel design, employing a highly integrated measurement scheme to acquire analog outputs from two bridge circuits. The circuitry for all five fingers is identical and interchangeable. The circuit board is mounted at the interphalangeal joint, flush against the crossbeam, facilitating easy disassembly and maintenance.

[0049] S30 acquires joint position information and fingertip torque information through magnetic coded sensors on finger joints and strain gauges on fingertips, and uses this information as data output for a force-position hybrid control method for dexterous hand fingers.

[0050] Please see Figure 5 As shown, in one embodiment of the present invention, after the magnetic encoder sensor and strain gauge are installed, initial values ​​need to be collected and saved to the internal FLASH of the MCU. In addition, the collected values ​​are low-pass filtered. The specific execution process is as follows: After system initialization, it checks whether there are any saved initial values ​​in FLASH. If there are, it reads the initial values ​​from FLASH and sets them, then enters the main processing loop. If not, it enters the initial value acquisition process: continuously acquires data from multiple sensors (referring to magnetic encoder sensors and strain gauges), calculates the average value, saves the average value to FLASH and sets the initial value, then enters the main processing loop.

[0051] The main processing loop is as follows: read the current sensor value, subtract the initial value, obtain the current difference, perform low-pass filtering on the current difference, output the processed value, wait for the sampling period, and then read the current sensor value again in the next period, and so on.

[0052] In this embodiment, joint position information and fingertip torque information are acquired through magnetically encoded sensors on the joints and strain gauges on the fingertips, which are used for data output in a force-position hybrid control method for dexterous hands. The design concept of the force-position hybrid control method uses a position controller as the inner loop to ensure tracking accuracy. In the outer loop, a position correction is calculated based on the measured force, allowing the finger to "retract" like a spring when it encounters an object, thus achieving compliant control. Please refer to [link to relevant documentation]. Figure 6 ,7 As shown, specifically: S31 simplifies the three joints of a dexterous hand's fingers into a three-bar linkage model.

[0053] In this embodiment, the current joint angle is first obtained from the magnetic coded sensors of the three joints. The fingertip contact force F_tip is obtained from the strain gauge. The three joints of the dexterous hand's fingers can be simplified into a three-bar model. Based on the angle information of the magnetically encoded sensors of the three joints and the equivalent link length, the current Cartesian space position of the fingertip is calculated through geometric forward kinematics.

[0054] S32 obtains the current fingertip position based on the angle information of the magnetic coding sensors on the three joints and the equivalent link length.

[0055] In this embodiment, obtaining the current fingertip position includes: S321 models the motion position and posture of the bending degree of a single finger of a dexterous hand. For the structure of a single finger, a single finger joint coordinate system is established to obtain the finger structure parameters of each joint segment.

[0056] In this embodiment, DH parameters are used to model and analyze the motion position and posture of the bending degree of freedom of a single finger in a dexterous hand. For the structure of a single finger, a model is established as follows: Figure 6 The single-knuckle coordinate system shown uses a systematic coordinate transformation to calculate the position coordinates of the fingertip in three-dimensional space, and then uses this as a basis to further determine its working space range.

[0057] S322, substitute the finger structure parameters of each joint segment into the DH coordinate and change them to obtain the pose change matrix of each finger joint.

[0058] In this embodiment, please refer to Table 1 for the finger DH parameters.

[0059] Table 1 Finger DH Parameter Table

[0060] Among them, combined with the appendix Figure 6 In the table i This is the linkage designation for the finger mechanism. This refers to the rotation around the x-axis during the DH coordinate transformation, since the finger joints are parallel. All are 0. This refers to the movement along the x-axis during the DH coordinate transformation, therefore , , These are the lengths of the corresponding finger root segment, finger middle segment, and fingertip segment, respectively. In the DH coordinate transformation, the movement along the z-axis occurs with the dexterous hand's fingers moving vertically upwards without any lateral movement. d All are 0. To allow for the rotation angle of the dexterous hand joints. , , These are the rotation angles of the metacarpophalangeal joint (MCP), proximal interphalangeal joint (PIP), and distal interphalangeal joint (DIP), respectively.

[0061] Substituting the structural parameters of each finger segment into the general formula of DH transformation, the pose transformation matrix of each finger joint is obtained. : ; In the formula, s represents the sine function sin and c represents the cosine function cos.

[0062] S323, according to the chain rule of coordinate system transformation, the pose transformation matrix from the fingertip coordinate system to the finger root base coordinate system can be obtained.

[0063] In this embodiment, the fingertip coordinate system can be obtained according to the chain rule of coordinate system transformation. To the root coordinates pose transformation matrix : ; ; S324: Obtain the current position of the fingertip based on the pose transformation matrix from the fingertip coordinate system to the finger root base coordinate system.

[0064] In this embodiment, the current position of the fingertip is represented by the following formula: ; ; In the formula, The position of the front fingertip, , , The lengths of the base, middle, and tip of the fingers, respectively. , , Let θ represent the rotation angles from the MCP, PIP, and DIP joints, respectively, where cos is the cosine function and sin is the sine function.

[0065] S33 calculates the current fingertip speed by performing numerical difference on historical values ​​of fingertip position information.

[0066] S34 sets a virtual spring and damping, and obtains the control correction amount based on the current fingertip speed and fingertip torque information.

[0067] In this embodiment, a virtual spring and damper are defined: ; In the formula, This represents virtual stiffness. The smaller the value, the softer the control effect of the finger, meaning the greater the influence of the contact force. It is a virtual damper used to absorb energy and prevent system oscillation. The current fingertip speed is obtained through kinematic calculation and differentiation using a magnetically encoded sensor. This is the calculated control correction amount.

[0068] S35, based on the control correction amount for the original fingertip Perform offset correction to obtain a new, compliant target position.

[0069] In this embodiment, the formula for obtaining the new, compliant target position is: ; Original fingertip position Use control correction amount After offset correction, a new, compliant target position is obtained. .

[0070] S36, the target values ​​of joint angles for each joint are obtained by inverse kinematics calculation of the new, compliant target position.

[0071] S37 uses a PID controller to control the joint motor actuator based on the target joint angle values ​​of each joint.

[0072] In this embodiment, the target joint angle value θ_soft_target for each joint is finally obtained through inverse kinematics calculation, and the PID controller reaches the specified target position. The specific method of inverse kinematics calculation is not limited in this invention. Through this design, the finger accurately tracks the original target position when there is no contact force. When force is applied, the finger automatically deviates from the original trajectory, with the offset proportional to the force, behaving like a spring. The method is simple, requires relatively little computing power, and can be easily deployed on the end effector. The data processing of the sensing system is also very simple. The force-position hybrid control system constructed in this way is simple and efficient, requiring only a common microcontroller control system, greatly reducing the implementation threshold of a dexterous hand compliant control system.

[0073] In the algorithm, the virtual stiffness can be dynamically adjusted according to the characteristics of the object being grasped. The smaller the value, the softer the control the fingers exhibit, meaning the higher the flexibility.

[0074] Please see Figure 8 As shown, the present invention also provides a system applying the above-described force-position hybrid five-finger dexterity hand-feel control method, comprising: The MCU master controller is used to acquire joint position information and fingertip torque information, and as a calculation unit, it acquires data output for the force-position hybrid control method.

[0075] The joint motor actuator is connected to the MCU master controller and is used to execute the data output by the force-position hybrid control method according to the control of the MCU master controller.

[0076] In this embodiment, the system also includes an IIC expansion chip for connecting to the acquisition circuits of 15 magnetically encoded sensors and 10 strain gauges. After acquiring data, the data is transmitted to the MCU main controller. It can be seen that the MCU main controller, as a data acquisition and processing unit, uses the IIC expansion chip to acquire data from the 15 magnetically encoded sensors and 10 strain gauges. Simultaneously, the MCU main controller, as a calculation unit, calculates the data from the magnetically encoded sensors and strain gauges, and implements compliant force-position mixing control of the finger.

[0077] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention, and no reference numerals in the claims should be construed as limiting the scope of the claims.

[0078] The above embodiments are merely examples of implementation methods of the invention. The scope of protection of the present invention is not limited to the above embodiments. For those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all fall within the scope of protection of the present invention.

Claims

1. A force-position hybrid five-finger dexterity tactile control method, characterized in that, include: The magnetic coding sensor adopts a modular design scheme, integrating the magnetic coding sensors of three joints on each finger onto a flexible circuit board; Two strain gauges are attached to the fingertip joint in the horizontal direction and two in the vertical direction to form a two-way stress measurement bridge. The joint position information and fingertip torque information are obtained by magnetic coding sensors on the finger joints and strain gauges on the fingertips, which are used for the data output of the force-position hybrid control method for dexterous hand fingers.

2. The force-position hybrid five-finger dexterity hand-feel control method according to claim 1, characterized in that, The flexible circuit board includes a slender main circuit flexible board (10), a first slender branch circuit flexible board (11), and a second slender branch circuit flexible board (12); the magnetic coding sensors set from the fingertip joint and the two joints below the fingertip joint are defined as the first magnetic coding sensor (21), the second magnetic coding sensor (22), and the third magnetic coding sensor (23) in sequence. Among them, a slender main circuit flexible board (10) is attached along the surface of the bionic finger, and one end of the slender main circuit flexible board (10) is defined as the main end (101) and the other end as the main root (102); a first magnetic coding sensor (21) is disposed on the main end (101) and attached to the fingertip joint of the bionic finger, and a gold finger interface (13) is provided on the main root (102) of the slender main circuit flexible board (10). The first slender branch circuit flexible plate (11) and the second slender branch circuit flexible plate (12) are arranged along both sides of the slender main circuit flexible plate (10); and the second magnetic coding sensor (22) and the third magnetic coding sensor (23) are respectively arranged at the ends of the first slender branch circuit flexible plate (11) and the second slender branch circuit flexible plate (12), and respectively attached to the two remaining joints of the bionic finger except for the fingertip joint; Electrical wiring between the three magnetically encoded sensors and the gold finger interface (13) is integrated on a flexible circuit board.

3. The force-position hybrid five-finger dexterity hand-feel control method according to claim 2, characterized in that, The flexible circuit board is locally reinforced at the fixed position of the magnetic encoder sensor to increase the thickness of the flexible circuit board.

4. The force-position hybrid five-finger dexterity hand-feel control method according to claim 2, characterized in that, The end of the main stem (101) is stepped.

5. The force-position hybrid five-finger dexterity hand-feel control method according to claim 1, characterized in that, The strain gauge is installed as a separate module at the fingertip joint.

6. The force-position hybrid five-finger dexterity hand-feel control method according to claim 1, characterized in that, The data output of the force-position hybrid control method for dexterous hand fingers includes: The three joints of the fingers of a dexterous hand are simplified into a three-bar linkage model; The current fingertip position is obtained based on the angle information of the magnetically encoded sensors on the three joints and the equivalent link length; The current fingertip speed is calculated by numerically differencing historical values ​​of fingertip position information. Define a virtual spring and damper, and obtain control correction values ​​based on the current fingertip speed and fingertip torque information; Based on the control correction amount for the original fingertip Perform offset correction to obtain a new, compliant target position; The target values ​​of joint angles for each joint are obtained by inverse kinematics calculation of the new, compliant target position. Based on the target joint angle values ​​of each joint, a PID controller is used to control the joint motor actuator.

7. The force-position hybrid five-finger dexterity hand-feel control method according to claim 6, characterized in that, Get the current fingertip position, including: Model the motion position and posture of the bending degree of freedom of a single finger in a dexterous hand. Establish a single finger joint coordinate system for the structure of a single finger and obtain the finger structure parameters of each joint segment. Substitute the finger structure parameters of each joint segment into the DH coordinate system and change them to obtain the pose change matrix of each phalanx. According to the chain rule of coordinate system transformation, the pose transformation matrix from the fingertip coordinate system to the finger root base coordinate system can be obtained. The current position of the fingertip is obtained based on the pose transformation matrix from the fingertip coordinate system to the finger root base coordinate system.

8. The force-position hybrid five-finger dexterity hand-feel control method according to claim 7, characterized in that, The current position of the fingertip is represented by the following formula: ； ； In the formula, The position of the front fingertip, , , These are the lengths of the three joint segments from the fingertip to the base of the finger. , , denoted as the rotation angles of the three joint segments from the fingertip to the base of the finger around the Z-axis, where cos is the cosine function and sin is the sine function.

9. The force-position hybrid five-finger dexterity hand-feel control method according to claim 6, characterized in that, The new, compliant target position is obtained using the following formula: ； ； In the formula, For the new, compliant target location, For the original fingertips , To control the amount of correction, For virtual stiffness, This is information about fingertip torque. For virtual damping, This represents the current fingertip speed.

10. A system applying the force-position hybrid five-finger dexterity hand-feel control method according to any one of claims 1-9, characterized in that, include: The MCU main controller is used to acquire joint position information and fingertip torque information, and as a calculation unit, it acquires data output for the force-position hybrid control method. The joint motor actuator is connected to the MCU master controller and is used to execute the data output by the force-position hybrid control method according to the control of the MCU master controller.

Images