Area stiffness dispensing and gradient transition bionic dexterous hand based on contact sensing

Abstract

This invention provides a bionic dexterous hand based on contact perception, featuring regional stiffness allocation and gradient transition. The dexterous hand includes a palm base, multiple bionic fingers, a tactile sensing component, a regional stiffness adjustment component, and a control component. Along the contact transmission path from the distal fingertips to the palm, a high-stiffness fingertip manipulation area, a mid-section gradient transition area, and a proximal low-stiffness covering area are sequentially arranged. The high-stiffness area contains a rigid skeleton and locking elements; the low-stiffness area consists of an elastic mesh and a flexible covering layer; and the gradient transition area achieves continuous stiffness variation through tendon ligaments and a pre-tension adjustment actuator. The control component determines the current task as fine manipulation, covering grasp, or a combined manipulation mode based on tactile perception indicators. In the combined manipulation mode, the system maintains high stiffness in the high-stiffness area while adjusting the tendon ligament pre-tension, creating a continuous stiffness distribution from high to low along the contact path in the gradient transition area, thus balancing fine manipulation, compliant covering, and combined grasping stability.

Description

Technical Field

[0001] This invention relates to the field of robotics, and more specifically, to a biomimetic dexterous hand based on contact perception, with region stiffness assignable and gradient transition. Background Technology

[0002] Bionic dexterous hands are important end effectors for service robots, industrial collaborative robots, remotely operated robots, and medical assistive devices. Existing dexterous hands typically need to simultaneously meet two conflicting functional requirements: one is the high rigidity, high positioning accuracy, and high force transmission efficiency required when the fingertips make point contact, pinch, twist, or press on small targets, slender objects, or function buttons; the other is the low rigidity, compliant fit, and large contact area required when performing surface contact, covering, and grasping of targets with varying sizes and irregular surface features, such as cups, fruits and vegetables, and daily necessities.

[0003] To address the aforementioned conflicting needs, existing technologies are mainly improved along the following paths: First, by increasing the degrees of freedom of finger joints and using underactuated or coupled transmission mechanisms to enhance shape adaptability; second, by changing the overall stiffness of the hand or fingers through magnetorheological materials, particle blocking, elastic tendon pretensioning, flexible beams, or other variable stiffness units; and third, by improving local performance through deformable palms, tactile fingertips, or methods based on driving current-contact pressure mapping.

[0004] However, existing technologies have at least the following shortcomings. First, most solutions primarily adjust stiffness uniformly across the entire hand or independently at a single joint, lacking a technical approach that collaboratively partitions the distal fingertip, intermediate phalanges, proximal phalanges, and palm according to contact function along the same effective contact transmission path. This makes it difficult to simultaneously achieve high local stiffness at the fingertip and low contact stiffness on the palm surface in the same task. Second, stiffness adjustment in many solutions still exhibits abrupt switching or boundary changes, lacking a continuous and gradual gradient transition between regions. This easily leads to stress concentration, impact, and vibration during contact establishment, load transfer, or slip suppression stages. Third, existing sensing solutions typically focus on single pressure values, end-effector force sensing, or image-assisted classification, failing to fully utilize multi-feature fusion information such as contact area, region proportion, pressure dispersion, contact center migration, and contact force change rate to classify operating modes and drive real-time reconstruction of regional stiffness. Fourth, existing rigid-flexible coupling structures mostly remain at the level of the finger's local or overall skeleton, rarely constructing a continuous covering contact area together with the proximal phalanges and adjacent palm.

[0005] Therefore, a new bionic dexterous hand technology solution is needed to form a functional continuum at the mechanical structure level, consisting of a "high-stiffness fingertip manipulation area - gradient transition area - low-stiffness covering area". The low-stiffness covering area spans the proximal phalanx and the adjacent palm, while the gradient transition area is composed of a flexible connecting end, a flexible beam, tendons, and a pre-tensioning adjustment actuator. At the perception level, multi-dimensional contact features are extracted, and at the control level, pattern determination and zoned collaborative control laws are established to solve the problem of the difficulty in simultaneously achieving fine manipulation and compliant grasping.

[0006] Patent application CN121179455A discloses a rigid-soft coupled bionic dexterous hand, including a dexterous hand mechanism, a wire-driven mechanism, a wireless control component, and a detachable fingertip flexible sensor. The hand mechanism adopts a "mirror thumb" configuration with two thumbs and three fingers, assembled into a palm groove by bolts. The joints of the thumb and fingers are integrally formed into elastic muscles after being molded and cured with liquid silicone rubber, achieving flexible connection and automatic reset of rigid joints. Drive ropes are threaded through oblique holes in the joints, and servo motors independently control the thumb's dual-degree-of-freedom lateral deviation / flexion and the fingers' single-degree-of-freedom movement. The fingertip flexible sensor is a flexible pressure sensor encapsulated in a silicone rubber film, which can be independently installed and removed, and the tactile signal is transmitted to an LED display for real-time feedback via the wireless control component. However, this patent cannot completely solve the existing technical problems, nor can it meet the needs of this invention. Summary of the Invention

[0007] To address the shortcomings of existing technologies, the purpose of this invention is to provide a biomimetic dexterous hand based on contact-sensing regional stiffness allocation and gradient transition.

[0008] The bionic dexterous hand based on contact sensing with regional stiffness allocation and gradient transition provided by the present invention includes: a palm base, multiple bionic fingers, a tactile sensing component, a regional stiffness adjustment component, and a control component; The multiple bionic fingers are connected to the palm base; The tactile sensing components are distributed on the contact surfaces of the fingertips, finger pads, and palm. The regional stiffness adjustment component includes, in sequence, a high-stiffness fingertip operation area, a gradient transition area, and a low-stiffness covering area along the effective contact transmission path of each bionic finger from the distal fingertip contact surface, through the middle phalanx contact surface, to the proximal phalanx and the adjacent palm contact surface. The high-rigidity fingertip operating area is disposed on the distal phalanx and / or fingertip contact surface, and includes a rigid frame and a switchable locking element; the rigid frame is used to provide structural support, and the switchable locking element is connected to the rigid frame to change the equivalent stiffness of the high-rigidity fingertip operating area by locking or releasing. The low-stiffness covering area includes at least the proximal phalanx contact surface and the adjacent palm base contact surface, including an elastic support grid and a flexible covering layer; the elastic support grid is disposed inside the flexible covering layer or combined with the flexible covering layer to provide passive shape conforming capability; The gradient transition zone is located in the middle phalanx and / or in the path segment connecting the high-stiffness fingertip operating area and the low-stiffness covering area, and includes a flexible connecting end, a flexible beam, a tendon cord coupled to the flexible beam, and a pretension adjustment actuator; the flexible connecting end is used to connect the gradient transition zone to the adjacent segment, the flexible beam is used to bear bending loads, the tendon cord extends along the surface or interior of the flexible beam and is connected to the pretension adjustment actuator, and the pretension adjustment actuator is used to change the pretension force of the tendon cord to adjust the equivalent stiffness of the gradient transition zone; The tactile sensing component is used to acquire at least one of the following in real time: contact position, contact area, local pressure, pressure distribution, contact center, contact force change rate, and vibration response; the sensitive unit of the tactile sensing component establishes a preset area mapping relationship with the high-stiffness fingertip operation area, gradient transition area, and low-stiffness covering area, and transmits the acquired tactile data to the control component; The control component is electrically or signalally connected to the tactile sensing component and the area stiffness adjustment component, respectively, for receiving tactile array data, performing filtering, threshold segmentation and area aggregation according to the area mapping relationship, and obtaining at least four of the following: high stiffness fingertip operation area ratio, low stiffness covering area ratio, normalized contact area, pressure dispersion, contact center migration, contact force change rate and vibration index. Then, the current operation mode is determined according to the threshold relationship. In the composite operation mode, the switchable locking element controls the high stiffness fingertip operation area to maintain high stiffness, the elastic support grid and flexible covering layer control the low stiffness covering area to maintain low stiffness, and the pre-tightening adjustment actuator is adjusted to make the gradient transition area form a continuously monotonically changing stiffness distribution along the effective contact transmission path.

[0009] Preferably, the control component receives tactile array data within the same sampling window and obtains at least four of the following: the area ratio of the high-stiffness fingertip operating area rt, the area ratio of the low-stiffness covering area rp, the normalized contact area Acn, the pressure dispersion Dpn, the contact center migration Dcn, the contact force change rate dFcn, and the vibration index avn. The total effective contact area Ac is obtained by summing all sensitive unit surfaces that are determined to be in effective contact:

[0010] The total contact force Fc is obtained by multiplying and summing the local pressure and area of ​​each effective contact element:

[0011] The area ratio rt of the high-rigidity fingertip operating area is:

[0012] The area ratio rp of the low-stiffness cladding region is:

[0013] The pressure dispersion Dp is: hour, ; in, ; The normalized contact area Acn is:

[0014] The normalized pressure dispersion Dpn is:

[0015] The rate of change of contact force dFc(k) is:

[0016] The normalized rate of change of contact force dFcn is:

[0017] The vibration index av is:

[0018] The normalized vibration index avn is:

[0019] in, For the first The contact validity indicator for each sensitive unit, when contact is established. ,otherwise ; For the first Area elements corresponding to each sensitive unit; For the first Local pressure values ​​of each sensitive unit; and These represent the sets of effective contact units within the high-stiffness fingertip operating area and the low-stiffness covering area, respectively. The sampling period; This is for absolute value operations; This refers to the total sensing coverage area or the mission-calibrated area. To determine the upper limit of the pressure dispersion of the calibrated sample, To calibrate the upper limit of the rate of change of contact force, This represents the number of vibration sampling points. For the first One vibration sample value, To calibrate the upper limit of the vibration index, For average contact pressure, Let be the total contact force at the k-th sampling time.

[0020] Preferably, the control component is further configured to: when , and At that time, determine the fine operation indication quantity. ;otherwise ; when and At that time, determine the amount of the grasp instruction. ;otherwise ; when , and When determining the composite operation indication quantity ;otherwise ; when When, it is determined to be a composite operation mode; when and When, it is determined to be in fine operation mode; when and When, it is determined to be an enveloping capture mode; when , and At this time, maintain the previous stable control state or switch to the default compliant standby state; in, , ,as well as These are the statistical boundaries used to characterize the following conditions under the fine operation mode: a high proportion of remote local contact, a small total contact area, and a relatively concentrated or discrete pressure distribution. "Higher" means above the preset high threshold, and "lower" means below the preset low threshold. and These are the statistical boundaries used to characterize the high contact ratio and large total contact area of ​​the proximal and palm regions in the enveloping grasping mode. , ,as well as These are used to characterize the simultaneous effective parameters of the distal and proximal ends in the combined operation mode, and the statistical boundary when in contact.

[0021] Preferably, coordinates are established along the effective contact path from a single finger to the palm. ,in Corresponding to the boundary of the fingertip operation area, Corresponding to the boundary of the enclosing region; in the composite operation mode, the control component makes the target stiffness function at any position x as:

[0022] in, The target stiffness value for the high-stiffness fingertip operating area. The target stiffness value for the low-stiffness cladding region, and ; The position coordinates defined along the effective contact transfer path. This represents the path length corresponding to the gradient transition region. Let x be a shape function whose control stiffness smoothly varies from high to low. , And in and The first derivatives at each point are all 0.

[0023] Preferably, the shape function Represented as: .

[0024] Preferably, the gradient transition region is composed of N discrete elastic elements, and the target stiffness of the j-th element is:

[0025] in, , where g is the gradient shape factor.

[0026] Preferably, the control component employs first-order inertial closed-loop adjustment to regulate the actual stiffness of each region:

[0027] And a stiffness change rate limit is added:

[0028] in, For the first The actual stiffness of each functional area For the first Target stiffness of each functional area; Pick These correspond to the high-stiffness fingertip operation area, the gradient transition area, and the low-stiffness covering area, respectively. This refers to the adjustment time constant for the corresponding region; This represents the maximum allowable rate of stiffness change without additional perturbation constraints. , For adjustment coefficients, To normalize the rate of change of contact force, For normalized vibration index, This is for absolute value operations, where t is the unit of time.

[0029] Preferably, the pretension adjustment actuator adjusts the equivalent stiffness of the gradient transition zone by changing the pretension force T of the tendon rope. satisfy:

[0030] in, For the gradient transition zone under preload The equivalent stiffness is as follows. The reference bending stiffness of the flexible beam. For chordae tendon pretension, This is the proportionality coefficient between the chord preload and the stiffness increment. It is an experience index. .

[0031] Preferably, the tactile sensing component includes a fingertip sensing array, a fingertip sensing array, and a palm sensing array; the fingertip sensing array is used to sense local high-pressure point contact and slippage precursors, and the fingertip sensing array and palm sensing array are used to sense the surface contact range, pressure distribution, and coverage state; each sensing unit outputs a local pressure value to the control component.

[0032] Preferably, the switchable locking element in the high-rigidity fingertip operating area adopts electromagnetic locking, friction locking, pawl locking, or particle blocking locking; the elastic support mesh of the low-rigidity covering area is made of TPU or silicone rubber.

[0033] Compared with the prior art, the present invention has the following beneficial effects: (1) Stiffness allocation is based on functional areas, so that the high-precision operation capability of the fingertips and the soft covering capability of the palm can coexist in the same dexterous hand system. (2) A continuous or quasi-continuous stiffness distribution is constructed through a gradient transition zone to avoid the impact, vibration and stress concentration caused by traditional step stiffness adjustment; (3) By fusing multiple features such as contact area, region proportion, pressure dispersion, contact center migration and contact force change rate, the accuracy and robustness of operation mode recognition are improved; (4) Improve the grasping stability, vulnerability-friendly properties and composite operation adaptability in complex scenarios without significantly increasing the driving degrees of freedom. Attached Figure Description

[0034] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the overall structure of the bionic dexterous hand of the present invention.

[0035] Figure 2 This is a flowchart of the contact feature-based mode determination and stiffness control process of the present invention.

[0036] Figure 3 This is a schematic diagram of the continuous distribution function of the target stiffness of the present invention.

[0037] Figure 4 This is a schematic diagram of the tendon cord-flexible beam transition zone structure based on continuous adjustment of pretension force in Example 1 of the present invention.

[0038] Figure 5 This is a schematic diagram of the gradient transition region formed by multi-level elastic units in Example 2 of the present invention.

[0039] Figure labeling: 1-Palm base; 2-Index finger; 3-Middle finger; 4-Ring finger; 5-Little finger; 6-Thumb; 7-Control component; 8-Tactile sensing component; 9-Low stiffness covering area; 10-Gradient transition area; 11-High stiffness fingertip operating area; 12-Fingertip sensor array; 13-Fingertip sensor array; 14-Palm sensor array; 15-Flexible connection end; 16-Tendon chord; 17-Flexible beam; 18-Pretension adjustment actuator. Detailed Implementation

[0040] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0041] Example 1 This invention provides a bionic dexterous hand based on contact sensing, featuring regional stiffness allocation and gradient transition. It includes a palm base 1 and multiple bionic fingers connected to the palm base 1. The multiple bionic fingers include an index finger 2, middle finger 3, ring finger 4, little finger 5, and thumb 6. A tactile sensing component 8, a regional stiffness adjustment component, and a control component 7 are distributed on the fingertips, finger pads, and palm contact surfaces. The regional stiffness adjustment component, along an effective contact transmission path for each bionic finger from the distal fingertip contact surface, through the middle phalanx contact surface, to the proximal phalanx and adjacent palm contact surfaces, sequentially includes a high-stiffness fingertip operating area 11, a gradient transition area 10, and a low-stiffness covering area 9, thereby forming a cooperative structure of high stiffness at the distal end, continuous transition in the middle section, and low stiffness at the proximal end along the same contact path.

[0042] The high-rigidity fingertip operating area 11 is located on the distal phalanx and / or fingertip contact surface, including a rigid skeleton and a switchable locking element, for providing stable local high-rigidity output when performing point contact, pinching, pressing, flicking or twisting; the low-rigidity covering area 9 includes at least the proximal phalanx contact surface and the adjacent palm base 1 contact surface, including an elastic support grid and a flexible covering layer, to form a continuous covering contact area across the anatomical boundary between the finger and the palm; the gradient transition area 10 is located on the middle phalanx and / or in the path segment connecting the high-rigidity fingertip operating area (11) and the low-rigidity covering area 9, including a flexible connecting end 15, a flexible beam 17, a tendon chord 16 coupled to the flexible beam 17 and a pre-tensioning adjustment actuator 18, for achieving a continuous change in stiffness from high to low in space.

[0043] The pretension adjustment actuator 18 changes the pretension force of the tendon rope 16, causing the equivalent stiffness of the gradient transition zone 10 to change continuously and monotonically between the high-stiffness fingertip operating zone 11 and the low-stiffness covering zone 9. The switchable locking element in the high-stiffness fingertip operating zone 11 can be an electromagnetic locking, friction locking, pawl locking, or particle blocking locking method.

[0044] The tactile sensing component 8 is used to acquire at least one of the following in real time: contact position, contact area, local pressure, pressure distribution, contact center, contact force change rate, and vibration response; the sensitive unit of the tactile sensing component establishes a preset region mapping relationship with the high-stiffness fingertip operation area 11, gradient transition area 10, and low-stiffness covering area 9; in one embodiment, the tactile sensing component 8 includes a fingertip sensing array 12, a fingertip sensing array 13, and a palm sensing array 14.

[0045] The control component 7 is used to: receive tactile array data within the same sampling window, perform filtering, threshold segmentation, and region aggregation based on the region mapping relationship, and obtain the area ratio of the high-rigidity fingertip operation area 11. The low-stiffness covering area accounts for 9% of the total area. Normalized contact area Pressure dispersion Contact center migration Contact force change rate and vibration index At least four of them; then determine the current operation mode based on the threshold relationship, and in the compound operation mode, simultaneously control the high stiffness fingertip operation area 11 to maintain high stiffness and the low stiffness covering area 9 to maintain low stiffness, while adjusting the pre-tightening actuator 18 to make the gradient transition area 10 form a continuously monotonically changing stiffness distribution along the effective contact transmission path.

[0046] This invention achieves the coexistence of high-stiffness distal fingertip operation area 11, continuous transition area 10 in the middle section, and low-stiffness covering area 9 spanning the proximal phalanx and adjacent palm on the same contact path by forming a collaborative structure. Combined with the area mapping, multi-feature pattern determination, and partitioned collaborative control under composite operation of the tactile sensing component 8, this invention enables the coexistence of high stiffness distal fingertip operation, low stiffness proximal fingertip operation, and continuous transition area in the middle section in the same task.

[0047] The control component receives tactile array data within the same sampling window and obtains at least four of the following: the area ratio of the high-stiffness fingertip operating area rt, the area ratio of the low-stiffness covering area rp, the normalized contact area Acn, the pressure dispersion Dpn, the contact center migration Dcn, the contact force change rate dFcn, and the vibration index avn. The fine-operation mode corresponds to contact characteristics with a high proportion of fingertip area, a small total contact area, and a relatively discrete pressure distribution. Its determination criteria can be expressed as:

[0048] The enveloping grasping mode corresponds to a contact characteristic with a high proportion of contact in the proximal and palm areas and a large total contact area. Its determination criteria can be expressed as follows:

[0049] When both the fine-grained operation feature and the encapsulation and grasping feature are valid within the same sampling window, it is determined to be a composite operation mode:

[0050] in, Average contact pressure; Total effective contact area The summation of all sensitive unit surfaces determined to be in effective contact can be expressed as:

[0051] Total contact force The result, obtained by multiplying and summing the local pressure and area of ​​each effective contact element, can be expressed as:

[0052] When effective contact is present, the average contact pressure It can be represented as:

[0053] When effective contact falls only within the three functional areas mentioned above, then:

[0054] Pressure dispersion The degree of dispersion of pressure distribution within the effective contact area can be expressed as: exist hour,

[0055] exist At that time, contact center satisfy:

[0056]

[0057]

[0058] For contact center The coordinates; Contact center migration The spatial change of the contact center between adjacent sampling times can be represented as:

[0059] Contact force change rate The rate of change of total contact force between adjacent sampling times can be expressed as:

[0060] The total contact force is sampled at time k; To facilitate comparison of unified thresholds, the normalized contact area is... Contact force with normalization It can be represented as:

[0061] Normalized mean pressure Normalized pressure dispersion and normalized rate of change of contact force It can be represented as:

[0062]

[0063]

[0064] Vibration index and its normalized vibration index Normalized quantity of contact center migration It can be represented as:

[0065] Where Ac is the total effective contact area, Fc is the total contact force, rt is the proportion of the high-stiffness fingertip operating area, and rp is the proportion of the low-stiffness covering area. The proportion of the gradient transition region, For the first The contact validity indicator for each sensitive unit, when contact is established. ,otherwise ; For the first Area elements corresponding to each sensitive unit; For the first Local pressure values ​​of each sensitive unit; , and These represent the sets of effective contact elements within the high-stiffness fingertip operating area, the gradient transition area, and the low-stiffness covering area, respectively. and The first The planar coordinates of each sensitive unit; For the first The coordinates of the contact center at each sampling time; The sampling period; This is for absolute value operations; It is the Euclidean norm. This refers to the total sensing coverage area or the mission-calibrated area. To calibrate the upper limit of contact force, To allow for an upper limit on work pressure, To determine the upper limit of the pressure dispersion of the calibrated sample, To calibrate the upper limit of the rate of change of contact force, This represents the number of vibration sampling points. For the first One vibration sample value, To calibrate the upper limit of the vibration index, A calibrated upper limit is set for the amount of contact center migration; in embodiments involving chord pretension adjustment, It is the proportionality coefficient between the chord preload and the stiffness increment.

[0066] In one embodiment, pattern determination employs a feature-based determination method with calibrated thresholds. These thresholds are not arbitrarily set, but are statistically obtained from tactile calibration data of representative fine manipulation task samples, covering grasping task samples, and composite manipulation task samples. They can be determined using quantile boundaries, minimum false positive rate criteria, ROC curve analysis, or empirical calibration methods with safety margins. Specifically, , and Statistical boundaries used to characterize "high proportion of remote local contact, small total contact area, and concentrated or discrete pressure distribution" under fine operation mode; and This is used to characterize the statistical boundary of "high contact ratio in the proximal and palm regions and large total contact area" in the enveloping grasping mode; , ,as well as These are used to characterize the simultaneous effective parameters of the distal and proximal ends in the combined operation mode, and the statistical boundary at contact. The above thresholds can be determined comprehensively through prototype repeat experiments, known task-labeled samples, and misclassification costs to ensure the repeatability and engineering feasibility of mode switching. Therefore, at the sampling time... : when , and At that time, determine the fine operation indication quantity. ;otherwise This criterion applies to situations where the proportion of distal local contact is high, the total contact area is small, and the pressure distribution is clearly differentiated.

[0067] when and At that time, determine the amount of the grasp instruction. ;otherwise This judgment formula corresponds to a grasping state in which the proximal and palm regions participate in sufficient contact and the overall contact area is large.

[0068] when , and When determining the composite operation indication quantity ;otherwise The pattern determination relationship is as follows: when When, it is determined to be a composite operation mode; when and When, it is determined to be in fine operation mode; when and When, it is determined to be an enveloping capture mode; when , and When necessary, the system maintains the previous stable control state or switches to the default compliant standby state. This decision chain ensures that the composite mode has higher priority, thereby avoiding mode switching conflicts.

[0069] Establish coordinates along the effective contact path from a single finger to the palm. ,in Corresponding to the boundary of the fingertip operation area, Corresponding to the boundary of the encapsulation region. To ensure that the gradient transition region forms a continuous monotonic stiffness distribution from high to low along the effective contact transfer path, at any position... The target stiffness function at that point can be expressed as:

[0070] The shape function in the above formula It can be represented as:

[0071] in, The target stiffness value for the high-stiffness fingertip operating area. The target stiffness value for the low-stiffness cladding region, and ; The position coordinates defined along the effective contact transfer path. This represents the path length corresponding to the gradient transition region. Let x be a shape function that smoothly controls the stiffness from high to low. The purpose of using this type of shape function is to ensure that the gradient transition region maintains zero slope near both the fingertip and the cladding edge, thereby avoiding abrupt changes in the stiffness curve at the boundaries and reducing impact and stress concentration during contact establishment, load transfer, and local slip suppression. This function satisfies: , And in and The first derivative at each point is 0 to avoid abrupt changes in stiffness slope at the boundary.

[0072] For a gradient transition region consisting of N discrete elements, the stiffness distribution of a continuous target can be discretized by using discrete elastic elements.

[0073] For by The gradient transition region is composed of discrete elastic elements, the first... The target stiffness of each element can be expressed as:

[0074]

[0075] in, This is the gradient shape factor, used to adjust the rate of change of the discrete stiffness distribution along the path direction; The actual stiffness adjustment process in each region can be implemented using a first-order inertial closed-loop method:

[0076] in, For the first The actual stiffness of each functional area For the first Target stiffness of each functional area Pick These correspond to the high-stiffness fingertip operation area, the gradient transition area, and the low-stiffness covering area, respectively. This represents the adjustment time constant for the corresponding region. The first-order inertial model described above is derived from the dynamic adjustment characteristics of actuators, locking components, and flexible members under finite response bandwidth, and can be used to describe the process of the actual stiffness of each region asymptotically converging towards the target stiffness. To suppress impact, a stiffness change rate constraint is further added:

[0077] in, This represents the maximum allowable rate of stiffness change without additional perturbation constraints. , For adjustment coefficients, To normalize the rate of change of contact force, This is a normalized vibration index. The basis of this constraint expression is that when the rate of change of contact force is large or the vibration index increases, the system is more prone to transient impacts, local slippage, or structural oscillations; therefore, the stiffness adjustment rate should be actively reduced. and When the stiffness is smaller, it allows for faster tracking of the target stiffness. In the composite operation mode, the target stiffness of the high-stiffness fingertip operation area is greater than the target stiffness of the gradient transition area along the path, and the target stiffness of the gradient transition area along the path is greater than the target stiffness of the low-stiffness covering area.

[0078] The above model does not restrict the unique values ​​of specific coefficients. Its key technical points are: using contact characteristics to drive mode determination, using mode determination to drive the target stiffness of the partition, and then using a continuous gradient function to constrain the spatial stiffness distribution of the transition zone. Its rationality is reflected in: and These respectively reflect the degree of involvement in distal and proximal contact. Reflecting the contact scale, Reflects pressure distribution characteristics, and These quantities respectively reflect the contact establishment and disturbance state; together they constitute observable criteria for "point contact fine operation", "surface contact enveloping grasping" and "composite contact task", and therefore can be directly used as the control basis for regional stiffness distribution and dynamic adjustment.

[0079] Example 2 like Figure 1 As shown, the bionic dexterous hand of the present invention includes an execution body composed of a palm base 1, an index finger 2, a middle finger 3, a ring finger 4, a little finger 5, and a thumb 6. A control component 7 and a tactile sensing component 8 are electrically or signal-connected to the execution body, respectively. Figure 1 Used to describe the overall structure of the system, without specifying the exact number of degrees of freedom or the driving layout.

[0080] Along the contact transmission path from a single finger to the palm, from the palm side to the fingertip side, it sequentially includes a low-stiffness covering area 9, a gradient transition area 10, and a high-stiffness fingertip operating area 11. This structure emphasizes the functional continuity and spatial adjacency along the contact path, where the low-stiffness covering area 9 mainly corresponds to the proximal phalanx and the adjacent palm contact surface, the gradient transition area 10 mainly corresponds to the middle phalanx region, and the high-stiffness fingertip operating area 11 mainly corresponds to the distal phalanx and the fingertip contact surface.

[0081] like Figure 1 As shown, the tactile sensing components can be divided into a fingertip sensing array 12, a fingertip sensing array 13, and a palm sensing array 14. The fingertip sensing array 12 is used to sense local high-pressure point contact and slippage precursors, while the fingertip sensing array 13 and the palm sensing array 14 are used to sense the surface contact range, pressure distribution, and coverage state.

[0082] like Figure 2 As shown, the control component first acquires the pressure matrix and contact position, then performs filtering, threshold segmentation, and feature extraction based on the region mapping relationship; subsequently, it calculates... , , , , and Equal feature quantities, and obtained through threshold relationships and Finally, the target stiffness and adjustment rate constraints for each region are output. This process can be implemented using an embedded processor, FPGA, or industrial controller.

[0083] like Figure 3 As shown, the target stiffness distribution function characterizes the continuous relationship between the target stiffness from the fingertip operating area to the coverage area and the contact path position. The figure shows a normalized schematic curve, which starts high and ends low, maintaining zero slope at both ends to reflect a smooth boundary transition.

[0084] In the fine operation mode, the high-stiffness fingertip operation area 11 maintains high stiffness, the gradient transition area 10 maintains medium stiffness, and the low-stiffness covering area 9 maintains low stiffness. In the covering and grasping mode, the low-stiffness covering area 9 and the gradient transition area 10 maintain relatively low stiffness, while the high-stiffness fingertip operation area 11 can be at basic stiffness or medium stiffness. In the composite operation mode, the high-stiffness fingertip operation area 11 maintains high stiffness, the low-stiffness covering area 9 maintains low stiffness, and the gradient transition area 10 undertakes the role of buffering and smooth load transfer.

[0085] Example 1: Figure 4The diagram illustrates the main implementation method based on continuous adjustment using a tendon-flexible beam. In this example, the dexterous hand includes a palm base 1 and five bionic fingers: index finger 2, middle finger 3, ring finger 4, little finger 5, and thumb 6. Each finger includes a proximal phalanx, a middle phalanx, and a distal phalanx. The distal phalanx tip and the fingertip covering layer together form a high-stiffness fingertip operating area 11, the middle phalanx forms a gradient transition area 10, and the proximal phalanx and its corresponding palm contact area together form a low-stiffness covering area 9. Figure 4 In the middle section, the flexible connection end 15 is used to compliantly connect the gradient transition zone with the adjacent section, and the tendon rope 16 is coupled to the flexible beam 17 and changes the equivalent stiffness of the gradient transition zone under the action of the pre-tightening adjustment actuator 18.

[0086] The high-rigidity fingertip operating area 11 is internally reinforced with an aluminum alloy or carbon fiber skeleton and equipped with a miniature electromagnetic locking component. Upon receiving a command for a fine-operation mode or a composite operation mode, the electromagnetic locking component locks the local rotational joint of the distal phalanx, thereby increasing the equivalent bending stiffness of the high-rigidity fingertip operating area 11 to [value missing]. When the lock is released, the high-rigidity fingertip operating area 11 retains basic compliance to avoid collision damage.

[0087] The low-stiffness encapsulation region 9 adopts a composite structure of elastic support mesh and silicone covering layer, and its equivalent stiffness is set to... The elastic support mesh can be made of TPU, silicone rubber, or other elastomers to provide passive shape conformation when gripping objects such as cups, fruits, and hammer handles.

[0088] The tactile sensing component 8 employs a flexible pressure array attached to the contact surfaces of the fingertips, fingertips, and palm, specifically including a fingertip sensing array 12, a fingertip sensing array 13, and a palm sensing array 14. Each sensitive unit can output a local pressure value. The control component 7 periodically acquires array data, with a sampling period of... The time can range from 2ms to 20ms. The original pressure matrix is ​​first subjected to median filtering and low-pass filtering, and then based on the threshold... Extract the effective contact area.

[0089] Let the pretension force of the tendon ligament be The reference bending stiffness of the flexible beam is If the chord constraint incremental stiffness and the preload satisfy a monotonically increasing relationship, then the equivalent stiffness of the gradient transition region can be expressed as follows: This relationship originates from the constraint enhancement mechanism of the chord-flexible beam coupling structure on the lateral deformation of the flexible beam under tension preload. As the preload increases, the equivalent bending stiffness of the flexible beam increases; therefore, a monotonically increasing function can be used to describe the mapping relationship between the preload and the equivalent stiffness.

[0090]

[0091] in, For the gradient transition zone under preload The equivalent stiffness is as follows. The reference bending stiffness of the flexible beam. For chordae tendon pretension, This is the proportionality coefficient between the chord preload and the stiffness increment. This is an experience index. Parameters and It can be obtained through static compression tests, bending tests, or finite element calibration; among which... Reflects the sensitivity of preload to stiffness increments. This reflects the degree of nonlinearity of the mapping relationship. When the structure's operating range approaches a linear response, There is clear engineering evidence in the vicinity.

[0092]

[0093] Within the near-linear working range with small deformation, it can be taken as... At this point, the equivalent stiffness and preload can be approximated as follows: This approximation is essentially a first-order linearization of the nonlinear stiffness mapping near the target operating point, which facilitates online calculation and real-time implementation by the controller.

[0094]

[0095] In a set of non-limiting calibrations, it can be Take the effective total coverage area of ​​the tactile array as the value, and... , , , , , , , , , and These are respectively taken as the upper limit of the prototype's permissible operation or the upper limit of the calibration sample. These values ​​can all be obtained through repeated prototype experiments, standard operating condition tests, and statistical analysis of task samples, and are not subjectively specified. When a detection... and When the control component 7 determines that it is in a composite operation mode, the relevant high-rigidity fingertip operation area 11 maintains high rigidity, the low-rigidity covering area 9 where the palm and near the knuckles are located maintains low rigidity, and the gradient transition area 10 adjusts the preload. Establish a continuous stiffness distribution. Taking the task of "gripping the hammer handle and pressing the button with the fingertip of index finger 2" as an example, the high-stiffness fingertip operation area 11 at the distal end of index finger 2 is controlled to a high-stiffness state, while the palm and the proximal ends of the other fingers maintain low stiffness to envelop the hammer handle. The middle gradient transition area 10 smoothly transmits the load from the fingertip to the palm, reducing the local impact at the moment of pressing the button.

[0096] Example 2: Figure 5 As shown, this is an alternative implementation of the gradient transition for multi-level elastic elements. The difference between Example 2 and Example 1 lies in the implementation of the gradient transition region 10. The gradient transition region 10 is composed of... to It consists of elastic units arranged in series or parallel, with each elastic unit having different structural dimensions, material modulus, or preload state, thus causing its target stiffness to change monotonically along the contact path.

[0097] For by The gradient transition region is composed of discrete units, the first... The target stiffness of each element can be expressed as:

[0098]

[0099] in, This is the gradient shape factor, used to adjust the rate of change of the discrete stiffness distribution along the path direction; At that time, the stiffness change is more gradual on the side closer to the fingertip; At that time, the stiffness change is more gradual on the side closer to the palm; At this time, it corresponds to an approximately linear transition. Parameters The stiffness distribution can be tuned according to the target object size, surface fragility, required load transfer path, and material distribution in the transition zone, so that the stiffness distribution of the discrete element combination can better approximate the continuous target curve. When the control component detects higher Smaller and When the value is large, a higher target stiffness is output to the elastic element closer to the fingertip; when a target stiffness is detected... higher and When the target stiffness is large, the unit closer to the palm maintains a lower target stiffness. In the compound operation mode, the unit closer to the fingertip maintains higher stiffness, the unit closer to the palm maintains lower stiffness, and the middle unit follows... A continuous transition is formed through discrete approximation to reduce abrupt changes at stiffness boundaries.

[0100] Those skilled in the art will understand that, in addition to implementing the system, apparatus, and their modules provided by this invention in purely computer-readable program code, the same program can be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system, apparatus, and their modules provided by this invention can be considered a hardware component, and the modules included therein for implementing various programs can also be considered structures within the hardware component; alternatively, modules for implementing various functions can be considered both software programs implementing the method and structures within the hardware component.

[0101] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A biomimetic dexterous hand based on contact-sensing, region-stiffness-assignable, and gradient-transition design, characterized in that... include: Hand base, multiple bionic fingers, tactile sensing components, area stiffness adjustment components, and control components; The multiple bionic fingers are connected to the palm base; The tactile sensing components are distributed on the contact surfaces of the fingertips, finger pads, and palm. The regional stiffness adjustment component includes, in sequence, a high-stiffness fingertip operation area, a gradient transition area, and a low-stiffness covering area along the effective contact transmission path of each bionic finger from the distal fingertip contact surface, through the middle phalanx contact surface, to the proximal phalanx and the adjacent palm contact surface. The high-rigidity fingertip operating area is disposed on the distal phalanx and / or fingertip contact surface, and includes a rigid frame and a switchable locking element; the rigid frame is used to provide structural support, and the switchable locking element is connected to the rigid frame to change the equivalent stiffness of the high-rigidity fingertip operating area by locking or releasing. The low-stiffness covering area includes at least the proximal phalanx contact surface and the adjacent palm base contact surface, including an elastic support grid and a flexible covering layer; the elastic support grid is disposed inside the flexible covering layer or combined with the flexible covering layer to provide passive shape conforming capability; The gradient transition zone is located in the middle phalanx and / or in the path segment connecting the high-stiffness fingertip operating area and the low-stiffness covering area, and includes a flexible connecting end, a flexible beam, a tendon cord coupled to the flexible beam, and a pretension adjustment actuator; the flexible connecting end is used to connect the gradient transition zone to the adjacent segment, the flexible beam is used to bear bending loads, the tendon cord extends along the surface or interior of the flexible beam and is connected to the pretension adjustment actuator, and the pretension adjustment actuator is used to change the pretension force of the tendon cord to adjust the equivalent stiffness of the gradient transition zone; The tactile sensing component is used to acquire at least one of the following in real time: contact position, contact area, local pressure, pressure distribution, contact center, contact force change rate, and vibration response; the sensitive unit of the tactile sensing component establishes a preset area mapping relationship with the high-stiffness fingertip operation area, gradient transition area, and low-stiffness covering area, and transmits the acquired tactile data to the control component; The control component is electrically or signalally connected to the tactile sensing component and the area stiffness adjustment component, respectively, for receiving tactile array data, performing filtering, threshold segmentation and area aggregation according to the area mapping relationship, and obtaining at least four of the following: high stiffness fingertip operation area ratio, low stiffness covering area ratio, normalized contact area, pressure dispersion, contact center migration, contact force change rate and vibration index. Then, the current operation mode is determined according to the threshold relationship. In the composite operation mode, the switchable locking element controls the high stiffness fingertip operation area to maintain high stiffness, the elastic support grid and flexible covering layer control the low stiffness covering area to maintain low stiffness, and the pre-tightening adjustment actuator is adjusted to make the gradient transition area form a continuously monotonically changing stiffness distribution along the effective contact transmission path.

2. The biomimetic dexterous hand based on contact sensing with region stiffness allocation and gradient transition as described in claim 1, characterized in that, The control component receives tactile array data within the same sampling window and obtains at least four of the following: the area ratio of the high-stiffness fingertip operating area rt, the area ratio of the low-stiffness covering area rp, the normalized contact area Acn, the pressure dispersion Dpn, the contact center migration Dcn, the contact force change rate dFcn, and the vibration index avn. The total effective contact area Ac is obtained by summing all sensitive unit surfaces that are determined to be in effective contact: The total contact force Fc is obtained by multiplying and summing the local pressure and area of ​​each effective contact element: The area ratio rt of the high-rigidity fingertip operating area is: The area ratio rp of the low-stiffness cladding region is: The pressure dispersion Dp is: hour, ; in, ; The normalized contact area Acn is: The normalized pressure dispersion Dpn is: The rate of change of contact force dFc(k) is: The normalized rate of change of contact force dFcn is: The vibration index av is: The normalized vibration index avn is: in, For the first The contact validity indicator for each sensitive unit, when contact is established. ,otherwise ; For the first Area elements corresponding to each sensitive unit; For the first Local pressure values ​​of each sensitive unit; and These represent the sets of effective contact units within the high-stiffness fingertip operating area and the low-stiffness covering area, respectively. The sampling period; This is for absolute value operations; This refers to the total sensing coverage area or the mission-calibrated area. To determine the upper limit of pressure dispersion of the calibrated sample, To calibrate the upper limit of the rate of change of contact force, This represents the number of vibration sampling points. For the first One vibration sample value, To calibrate the upper limit of vibration index, For average contact pressure, Let be the total contact force at the k-th sampling time.

3. The contact-aware based regional stiffness assign and gradient transition emulating dexterous hand of claim 2, wherein, The control component is also used for: When , and , determine the fine operation instruction amount ; otherwise ; When and , determine the amount of coated gripping instruction ; otherwise ; when , and When determining the composite operation indication quantity ;otherwise ; when When, it is determined to be a composite operation mode; when and When, it is determined to be in fine operation mode; when and When, it is determined to be an enveloping capture mode; when , and At this time, maintain the previous stable control state or switch to the default compliant standby state; in, , ,as well as These are the statistical boundaries used to characterize the following conditions under the fine operation mode: a high proportion of remote local contact, a small total contact area, and a relatively concentrated or discrete pressure distribution. "Higher" means above the preset high threshold, and "lower" means below the preset low threshold. and These are the statistical boundaries used to characterize the high contact ratio and large total contact area of ​​the proximal and palm regions in the enveloping grasping mode. , ,as well as These are used to characterize the simultaneous effective parameters of the distal and proximal ends in the composite operation mode, and the statistical boundary at contact; k represents the kth sampling time.

4. The biomimetic dexterous hand based on contact sensing with region stiffness allocation and gradient transition as described in claim 1, characterized in that, Establish coordinates along the effective contact path from a single finger to the palm. ,in Corresponding to the boundary of the fingertip operation area, Corresponding to the boundary of the enclosing region; in the composite operation mode, the control component makes the target stiffness function at any position x as: in, The target stiffness value for the high-stiffness fingertip operating area. The target stiffness value for the low-stiffness cladding region, and ; The position coordinates defined along the effective contact transfer path. This represents the path length corresponding to the gradient transition region. Let x be a shape function whose control stiffness smoothly varies from high to low. , And in and The first derivatives at each point are all 0.

5. The biomimetic dexterous hand based on contact sensing with region stiffness allocation and gradient transition according to claim 4, characterized in that, The shape function Represented as: .

6. The biomimetic dexterous hand based on contact sensing with region stiffness allocation and gradient transition according to claim 1, characterized in that, The gradient transition region is composed of N discrete elastic elements, and the target stiffness of the j-th element is: in, , where g is the gradient shape factor.

7. The biomimetic dexterous hand based on contact sensing with assignable regional stiffness and gradient transition as described in claim 1, characterized in that, The control component uses first-order inertial closed-loop adjustment to regulate the actual stiffness of each region: And a stiffness change rate limit is added: in, For the first The actual stiffness of each functional area For the first Target stiffness of each functional area; Pick These correspond to the high-stiffness fingertip operation area, the gradient transition area, and the low-stiffness covering area, respectively. This refers to the adjustment time constant for the corresponding region; This represents the maximum allowable rate of stiffness change without additional perturbation constraints. , For adjustment coefficients, To normalize the rate of change of contact force, For normalized vibration index, This is for absolute value operations, where t is the unit of time.

8. The biomimetic dexterous hand based on contact sensing with assignable regional stiffness and gradient transition according to claim 1, characterized in that, The pretension adjustment actuator changes the pretension force T of the tendon rope to adjust the equivalent stiffness of the gradient transition zone. satisfy: in, For the gradient transition zone under preload The equivalent stiffness is as follows. The reference bending stiffness of the flexible beam. For chordae tendon pretension, This is the proportionality coefficient between the chord preload and the stiffness increment. It is an experience index. .

9. The biomimetic dexterous hand based on contact sensing with region stiffness allocation and gradient transition according to claim 1, characterized in that, The tactile sensing component includes a fingertip sensing array, a fingertip sensing array, and a palm sensing array; the fingertip sensing array is used to sense local high-pressure point contact and slippage precursors, while the fingertip and palm sensing arrays are used to sense the surface contact range, pressure distribution, and coverage state; each sensing unit outputs a local pressure value to the control component.

10. The biomimetic dexterous hand based on contact sensing with assignable regional stiffness and gradient transition according to claim 1, characterized in that, The switchable locking mechanism in the high-rigidity fingertip operating area employs electromagnetic locking, friction locking, pawl locking, or particle blocking locking methods; the elastic support mesh in the low-rigidity covering area is made of TPU or silicone rubber.

Images