Robot grasping posture adaptive adjustment device based on fusion of vision and touch
Patent Information
- Application Number
- CN202610752933.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-28
- Publication Date
- 2026-08-28
AI Technical Summary
综上所述本发明人发现,现有的自适应调节装置主要存在以下缺陷:由于当前自适应调节装置进行对不同物件夹取时,其会因夹爪形状的不同而产生的形状不匹配情况,同时当前自适应调节装置的末端执行器所携带的夹爪形状为统一形状,无法根据实际物品的形状进行改变,因此会出现夹取物件的局限性,为此会导致当前装置无法对多种形状物件进行夹取,使之会降低了当前装置的实际使用广泛性及使用强度
1.本发明由多功能夹持结构改进后,通过装载结构的组装框内壁拼接轨及锁定栓能有效的与夹块进行快速装配,同时结合电磁块的通电或断电进行对夹块吸附固定或松开,使之能够达成快速拆装的效果,同时能够根据所夹持物件对夹块形状进行更换,防止了所形成的夹爪与物件形状不匹配产生的无法抓取,以至于能够提高了当中自适应调节装置的使用效果。
Smart Images

Figure CN122645285A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of adaptive adjustment technology, and more specifically to an adaptive adjustment device for robot grasping posture based on the fusion of vision and touch. Background Technology
[0002] The vision and tactile fusion robot grasping posture adaptive adjustment device is an adaptive adjustment intelligent system composed of an end effector, multimodal sensors, and intelligent control algorithms. It enables the grasping robot to adjust its movements in real time based on tactile feedback the moment it comes into contact with an object through the intelligent combination of the two types of sensory information, thereby performing more precise and intelligent operations. In summary, the inventors have found that existing adaptive adjustment devices have the following main drawbacks: when gripping different objects, the current adaptive adjustment device may experience shape mismatch due to the different shapes of the grippers. Furthermore, the gripper shape carried by the end effector of the current adaptive adjustment device is uniform and cannot be changed according to the actual shape of the object. Therefore, there are limitations in gripping objects, which means that the current device cannot grip objects of various shapes, thus reducing the practical applicability and intensity of use of the current device. Summary of the Invention
[0003] The technical solution adopted by the present invention to achieve the technical objective is: a robot grasping posture adaptive adjustment device based on vision and tactile fusion, the structure of which includes: a rotating groove, a rotating rod, an extender, an intelligent control host, a slide rail, and a multi-functional clamping structure. The rotating groove and the rotating rod are integrated, and one end of the rotating rod is connected to the extender. One end of the extender is fixedly electrically connected to the intelligent control host. The intelligent control host is movably electrically connected to the multi-functional clamping structure through the slide rail.
[0004] As a further improvement of the present invention, the multifunctional clamping structure is provided with a moving block, a connecting block, a vision camera, an auxiliary wheel, a loading structure, a clamping block, and a tactile sensor. The moving block is connected to the connecting block, and the vision camera mounted on the upper and lower positions of the connecting block is located at the edge of the loading structure and communicates with the tactile sensor of the clamping block. The auxiliary wheel is located at the edge of the connecting block.
[0005] As a further improvement of the present invention, the system formed by the overall components of the adaptive adjustment device consists of the following three major modules, forming a precise closed-loop control system: Module 1: Decision-making layer: Intelligent processing and control; Module 2: Execution Layer: Outputs precise actions; Module 3: Perception Layer: Acquiring multimodal information.
[0006] As a further improvement of the present invention, the decision layer controls the execution layer by controlling the position, force and torque of the control command. Then, the execution layer performs the grasping operation and provides real-time feedback to the perception layer. The visual, tactile and force data streams of the perception layer are fed back to the decision layer, forming a closed-loop control system.
[0007] As a further improvement of the present invention, the decision layer can perform multimodal data fusion corresponding to the intelligent control host, including AI and deep learning algorithms, and then perform adaptive policy generation, including diffusion models and reinforcement learning.
[0008] As a further improvement of the present invention, the execution layer can correspond to the moving block, loading structure, and clamping block of the multifunctional clamping structure, including a drive system, which is an electric servo motor, and the adaptive execution mechanism can correspond to the gripper formed by the moving block through the loading structure and the clamping block.
[0009] As a further improvement of the present invention, the perception layer can correspond to a visual camera, a tactile sensor, etc. Through the visual perception of the visual camera and the cooperation of the tactile sensor, the visual and tactile data streams can be fed back to the decision layer. At the same time, the force perception and body perception data located at the joint angle position of the rotating rod are also fed back to the decision layer, allowing the decision layer to intelligently control the execution layer and form an adaptive adjustment.
[0010] As a further improvement of the present invention, the loading structure is further provided with an energized structure. One end of the energized structure is equipped with an assembly frame and the inner wall of the assembly frame is provided with a splicing rail. A locking bolt is provided on the outer side of the splicing rail, i.e., the outer edge of the assembly frame. An electromagnetic block is also connected to the center of the assembly frame.
[0011] As a further improvement of the present invention, the energized structure is perpendicular to the assembly frame, the assembly frame is in the shape of a "U", the splicing rail of the assembly frame is set in a symmetrical orientation and is connected to the electromagnetic block, and the electromagnetic block is a raised solid shape.
[0012] As a further improvement of the present invention, the energized structure is further provided with a reinforcing block, one end of which is connected to an insulating frame, and a limiting frame is mounted on the inner wall of the insulating frame, and an electrical connection protrusion is included in the limiting frame.
[0013] As a further improvement of the present invention, the reinforcing block is a rubber product and there are two blocks arranged symmetrically at the upper and lower center positions of the insulating frame. The insulating frame and the limiting frame overlap each other and cover the edge of the electrical connection protrusion.
[0014] As a further improvement of the present invention, the extender is provided with a connecting end, the connecting end is integrated with the balance block, an overlapping layer is provided on the upper end of the balance block, a positioning block is provided at the center of the overlapping layer, a sliding rod is connected to the center of the positioning block, and an auxiliary rod is connected to the edge of the overlapping layer.
[0015] As a further improvement of the present invention, the balance block is fixedly connected to one end of the intelligent control host through the connecting end, the overlapping layer of the balance block is parallel to one end of the rotating rod, the slide rod includes a small drive motor and a rotating pulley, and there are four auxiliary rods distributed at the corners of the overlapping layer and inserted into the rotating rod.
[0016] As a further improvement of the present invention, the auxiliary rod is also provided with a locking rod, the upper end of the locking rod is connected to a bottom block, the upper center of the bottom block is provided with a rod body, and the upper end of the rod body is connected to an anti-detachment block.
[0017] As a further improvement of the present invention, the locking rod is provided with an external thread and forms a "T" shape with the bottom block. The upper end of the bottom block is perpendicular to the rod body, and the top area of the rod body is smaller than the area of the anti-detachment block.
[0018] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention improves upon the multi-functional clamping structure by using the splicing rails and locking bolts on the inner wall of the assembly frame of the loading structure to effectively and quickly assemble with the clamping block. Simultaneously, the clamping block is attracted, fixed, or released by the energization or de-energization of the electromagnetic block, achieving a rapid assembly and disassembly effect. Furthermore, the shape of the clamping block can be changed according to the object being clamped, preventing the inability to grip due to mismatch between the formed gripper and the object shape, thus improving the effectiveness of the adaptive adjustment device.
[0019] 2. The present invention improves the power-conducting structure mounted on the outer center of the assembly frame of the loading structure. The reinforcing block on the outside of the insulating frame can be reinforced and fixed in the insertion connecting block through its own rebound effect, which improves the stability of the electrical connection and prevents shaking and instability during gripping. Furthermore, the double protection of the limiting frame on the edge of the electrical connection protrusion and the insulating frame can prevent leakage at the clamping block and other locations caused by power leakage, thus improving the safety factor during use.
[0020] 3. This invention, with its improved extender, utilizes the overlapping layer of the balance block to ensure that the slide rod and auxiliary rod are parallel and aligned with the rotating rod after resetting inside. Conversely, the slide rod, driven by a built-in small drive motor, moves the pulley to slide out of the rotating rod, adjusting the position and height of the overall intelligent control host's multi-functional clamping structure. The combination of the small drive motor and pulley enables simpler and more convenient adjustment, simplifying control and avoiding the difficulties of subsequent maintenance caused by complex structures. Furthermore, during extension, the auxiliary rod at the corner of the overlapping layer provides balance and prevents swaying. The anti-derailment block on the auxiliary rod prevents derailment during slide rod extension, thus improving the overall usability of the device. Attached Figure Description
[0021] Figure 1 This is a structural diagram of a robot grasping posture adaptive adjustment device based on the fusion of vision and touch.
[0022] Figure 2 This is a three-dimensional structural diagram of an improved multi-functional clamping structure.
[0023] Figure 3 This is a schematic diagram of a closed-loop control system, which is a type of adaptive adjustment device.
[0024] Figure 4 This is a three-dimensional structural diagram of an improved loading structure.
[0025] Figure 5 This is a cross-sectional schematic diagram of an improved energized structure.
[0026] Figure 6 This is a three-dimensional structural diagram of an improved extender.
[0027] Figure 7 This is a schematic diagram of a three-dimensional structure of an improved auxiliary rod.
[0028] In the diagram: Rotary slot-1, Rotary rod-2, Extender-3, Intelligent control host-4, Slide rail-5, Multifunctional clamping structure-6; Moving block-61, connecting block-62, vision camera-63, auxiliary wheel-64, loading structure-65, clamping block-66, tactile sensor-67; Powered structure-651, assembly frame-652, splicing rail-653, locking bolt-654, electromagnetic block-655; Reinforcing block-6511, insulating frame-6512, limiting frame-6513, electrical connection protrusion-6514; Connecting end-31, balance block-32, overlapping layer-33, positioning block-34, slide bar-35, auxiliary bar-36; Locking lever-361, base block-362, lever body-363, anti-detachment block-364. Detailed Implementation
[0029] The present invention will be further described below with reference to the accompanying drawings: Example 1: Figures 1 to 5 As shown: This invention provides a robot grasping posture adaptive adjustment device based on vision and tactile fusion. Its structure includes: a rotating groove 1, a rotating rod 2, an extender 3, an intelligent control host 4, a slide rail 5, and a multi-functional clamping structure 6. The rotating groove 1 and the rotating rod 2 are integrated, and one end of the rotating rod 2 is connected to the extender 3. One end of the extender 3 is fixedly electrically connected to the intelligent control host 4. The intelligent control host 4 is movably electrically connected to the multi-functional clamping structure 6 through the slide rail 5.
[0030] The multifunctional clamping structure 6 includes a moving block 61, a connecting block 62, a vision camera 63, an auxiliary wheel 64, a loading structure 65, a clamping block 66, and a tactile sensor 67. The moving block 61 is connected to the connecting block 62, and the vision camera 63 mounted on the upper and lower positions of the connecting block 62 is located at the edge of the loading structure 65 and communicates with the tactile sensor 67 of the clamping block 66. The auxiliary wheel 64 is located at the edge of the connecting block 62.
[0031] The system formed by the overall components of the adaptive adjustment device consists of the following three major modules, forming a precise closed-loop control system: Module 1: Decision-making layer: Intelligent processing and control; Module 2: Execution Layer: Outputs precise actions; Module 3: Perception Layer: Acquiring multimodal information.
[0032] The decision-making layer controls the execution layer by controlling the position, force, and torque of the control commands. The execution layer then performs the grasping operation and provides real-time feedback to the perception layer. The visual, tactile, and force data streams from the perception layer are fed back to the decision-making layer, forming a closed-loop control system.
[0033] The decision layer can perform multimodal data fusion, including AI and deep learning algorithms, corresponding to the intelligent control host 4, and then perform adaptive policy generation, including diffusion models and reinforcement learning.
[0034] The execution layer can correspond to the moving block 61, loading structure 65, and clamping block 66 of the multi-functional clamping structure 6, and includes a drive system, which is an electric servo motor. The adaptive execution mechanism can correspond to the gripper formed by the moving block 61 through the loading structure 65 and the clamping block 66.
[0035] The perception layer can correspond to a visual camera 63, a tactile sensor 67, etc. Through the visual perception of the visual camera 63 and the cooperation of the tactile sensor 67, visual and tactile data streams can be fed back to the decision layer. At the same time, the force perception and body perception data located at the joint angle position of the rotating rod 2 are also fed back to the decision layer, allowing the decision layer to intelligently control the execution layer and form an adaptive adjustment.
[0036] The loading structure 65 is further provided with an energized structure 651. One end of the energized structure 651 is equipped with an assembly frame 652 and the inner wall of the assembly frame 652 is provided with a splicing rail 653. A locking bolt 654 is provided on the outer side of the splicing rail 653, i.e., the outer edge of the assembly frame 652. An electromagnetic block 655 is also connected to the center inside the assembly frame 652.
[0037] The energized structure 651 is perpendicular to the assembly frame 652. The assembly frame 652 is U-shaped. The splicing rail 653 of the assembly frame 652 is set in a symmetrical orientation and is connected to the electromagnetic block 655. The electromagnetic block 655 is a raised solid shape.
[0038] The energized structure 651 is further provided with a reinforcing block 6511. One end of the reinforcing block 6511 is connected to an insulating frame 6512. A limiting frame 6513 is mounted on the inner wall of the insulating frame 6512. The limiting frame 6513 contains an electrical connection protrusion 6514.
[0039] The reinforcing block 6511 is a rubber product, and there are two blocks arranged symmetrically at the upper and lower center positions of the insulating frame 6512. The insulating frame 6512 and the limiting frame 6513 overlap each other and cover the edge of the electrical connection protrusion 6514.
[0040] The specific functions and operation procedures of this embodiment are as follows: In this invention, the adaptive adjustment device can be connected to the rotation joint position of the robot base via the rotation slot 1 of the rotating rod 2. Then, the extender 3 at one end of the rotating rod 2 can adjust the position and height of the intelligent control host 4 and the multi-functional gripping structure 6, enabling stable gripping of objects. During gripping, the connecting block 62 of the multi-functional gripping structure 6 can slide in conjunction with the moving block 61 at the position of the slide rail 5, achieving a stable gripping effect. Then, combined with the vision camera 63 of the connecting block 62, the tactile sensor 67 on the gripping block 66 of the loading structure 65, and the intelligent control of the intelligent control host 4, the posture adaptive adjustment effect is achieved. It can form an integrated, precise, and complete closed-loop operation through the following process: Firstly, the system formed by the overall components of the adaptive adjustment device consists of the following three major modules, forming a precise closed-loop control system: Module 1: Decision-making layer: Intelligent processing and control; Module 2: Execution Layer: Outputs precise actions; Module 3: Perception Layer: Acquiring multimodal information; The decision-making layer controls the execution layer by controlling the position, force, and torque of the control commands. The execution layer then performs the grasping operation and provides real-time feedback to the perception layer. The visual, tactile, and force data streams from the perception layer are fed back to the decision-making layer, forming a closed-loop control system. The decision-making layer can perform multimodal data fusion, including AI and deep learning algorithms, in response to the intelligent control host 4, and then generate adaptive policies, including diffusion models and reinforcement learning. The execution layer can correspond to the moving block 61, loading structure 65, and clamping block 66 of the multi-functional clamping structure 6, including a drive system, which is an electric servo motor. The adaptive execution mechanism can correspond to the gripper formed by the moving block 61 through the loading structure 65 and the clamping block 66. The perception layer can correspond to the visual camera 63, tactile sensor 67, etc. Through the visual perception of the visual camera 63 and the cooperation of the tactile sensor 67, the visual and tactile data streams can be fed back to the decision layer. At the same time, the force and proprioception data located at the joint angle position of the rotating rod 2 are also fed back to the decision layer, allowing the decision layer to intelligently control the execution layer and form adaptive adjustment. This enables a stable gripping of objects. Furthermore, when dealing with objects of different shapes, the clamping block 66 can be quickly replaced via the assembly frame 652 of the loading structure 65. The energized structure 651 at one end of the assembly frame 652, combined with the reinforcing block 6511 and the insulating frame 6512, can be inserted into the connecting block 62 and then electrically connected to the intelligent control host 4. After insertion, the reinforcing block 6511 can eliminate assembly gaps through the rebound of the rubber components, forming a strong connection. Subsequently, the electrical connection protrusion 6514 in the limiting frame 6513 is tightly electrically connected to the inside of the connecting block 62. Simultaneously, the insulating frame 6512 prevents current leakage to other components, ensuring... To ensure safety during use, after the external locking bolt 654 of the assembly frame 652 is released, the edge of the clamp 66 of the splicing rail 653 can be released. At the same time, the electromagnetic block 655 is de-energized by the intelligent control host 4, thus achieving a demagnetizing effect. Then, the current clamp 66 can be pulled out to complete the disassembly. Conversely, the new clamp 66 with a tactile sensor 67 is pushed into the position of the splicing rail 653 until it interlocks with the electromagnetic block 655. Finally, the electromagnetic block 655 is energized again to "lock" the clamp 66 first, and then the locking bolt 654 is rotated to lock and fix the new clamp 66. This achieves the effect of quick disassembly and replacement, thereby improving the actual use strength of the overall adaptive adjustment device. Example 2: Figures 6 to 7 As shown: This invention provides a robot grasping posture adaptive adjustment device based on vision and tactile fusion. Its structure includes: the extender 3 is provided with a connecting end 31, the connecting end 31 is integrated with the balance block 32, the upper end of the balance block 32 is provided with an overlapping layer 33, the center of the overlapping layer 33 is provided with a positioning block 34, the center of the positioning block 34 is connected with a sliding rod 35, and an auxiliary rod 36 is connected to the edge of the overlapping layer 33.
[0041] The balance block 32 is fixedly connected to one end of the intelligent control host 4 through the connecting end 31. The overlapping layer 33 of the balance block 32 is parallel to one end of the rotating rod 2. The slide rod 35 includes a small drive motor and a rotating pulley. There are four auxiliary rods 36, which are distributed at the corners of the overlapping layer 33 and inserted into the rotating rod 2.
[0042] The auxiliary rod 36 is also provided with a locking rod 361. The upper end of the locking rod 361 is connected to a bottom block 362. The center of the upper end of the bottom block 362 is provided with a rod body 363. The upper end of the rod body 363 is connected to an anti-detachment block 364.
[0043] The locking rod 361 is provided with external threads and forms a "T" shape with the bottom block 362. The upper end of the bottom block 362 is perpendicular to the rod body 363. The top area of the rod body 363 is smaller than the area of the anti-detachment block 364.
[0044] The specific functions and operation procedures of this embodiment are as follows: In this invention, the balance block 32 of the extender 3 can be electrically connected to the center of one end of the intelligent control host 4 via the connecting end 31. Subsequently, the positioning block 34 of the upper overlapping layer 33 can determine the position of the slide rod 35, so that the slide rod 35 can replace the original complex moving structure according to the small electric motor and pulley it carries. This achieves a simpler and more convenient way to adjust the height of the multi-functional clamping structure 6, and also achieves the effect of easy maintenance in the future, further improving the ease of use of the extender 3. When adjusting the height, the corners of the overlapping layer 33 are set with... The four auxiliary rods 36 can achieve a position restraint effect. The bottom block 362 of the auxiliary rod 36 can be used to fix the rod body 363 vertically to one end corner of the balance block 32 through the locking rod 361. When the sliding rod 35 slides out from the inside of the rotating rod 2, it is driven to achieve an edge-limited auxiliary balance effect to prevent shaking. Furthermore, the path limitation of the anti-detachment block 364 at the upper end of the rod body 363 can prevent derailment and falling due to excessive sliding. This improves the stability of grasping objects and the protection of the height adjustment, avoiding excessive adjustment that would damage the actual safety factor of the overall device.
[0045] Any technical solution that achieves the above-mentioned technical effects by utilizing the technical solutions described in this invention, or by designing similar technical solutions by those skilled in the art under the inspiration of the technical solutions described in this invention, falls within the protection scope of this invention.
Claims
1. A robot grasping posture adaptive adjustment device based on vision and tactile fusion, the structure of which includes: The system comprises a rotating groove (1), a rotating rod (2), an extender (3), an intelligent control host (4), a slide rail (5), and a multi-functional clamping structure (6). The rotating groove (1) and the rotating rod (2) are integrated, with one end of the rotating rod (2) connected to the extender (3). One end of the extender (3) is fixedly electrically connected to the intelligent control host (4). The intelligent control host (4) is electrically connected to the multi-functional clamping structure (6) via the slide rail (5). The system is characterized by: The multifunctional clamping structure (6) is provided with a moving block (61), a connecting block (62), a vision camera (63), an auxiliary wheel (64), a loading structure (65), a clamping block (66), and a tactile sensor (67). The moving block (61) is connected to the connecting block (62), and the vision camera (63) mounted on the upper and lower positions of the connecting block (62) is located on the edge of the loading structure (65) and communicates with the tactile sensor (67) of the clamping block (66). The auxiliary wheel (64) is located on the edge of the connecting block (62).
2. The robot grasping posture adaptive adjustment device based on vision and tactile fusion according to claim 1, characterized in that: The system formed by the overall components of the adaptive adjustment device consists of the following three major modules, forming a precise closed-loop control system: Module 1: Decision-making layer: Intelligent processing and control; Module 2: Execution Layer: Outputs precise actions; Module 3: Perception Layer: Acquiring multimodal information; The decision-making layer controls the execution layer through control commands (position, force, torque), and then the execution layer performs the grasping operation while providing real-time feedback to the perception layer. The visual, tactile, and force data streams of the perception layer are fed back to the decision-making layer, forming a closed-loop control system.
3. The robot grasping posture adaptive adjustment device based on vision and tactile fusion according to claim 1, characterized in that: The decision layer can perform multimodal data fusion with the intelligent control host (4), including AI and deep learning algorithms, and then perform adaptive policy generation, including diffusion model and reinforcement learning.
4. The robot grasping posture adaptive adjustment device based on vision and tactile fusion according to claim 1, characterized in that: The execution layer can correspond to the moving block (61), loading structure (65), and clamping block (66) of the multi-functional clamping structure (6), which includes a drive system, which is an electric servo motor. The adaptive execution mechanism can correspond to the gripper formed by the moving block (61) through the loading structure (65) and the clamping block (66).
5. The robot grasping posture adaptive adjustment device based on vision and tactile fusion according to claim 1, characterized in that: The perception layer can correspond to a visual camera (63), a tactile sensor (67), etc. Through the visual perception of the visual camera (63) and the cooperation of the tactile sensor (67), the visual and tactile data streams can be fed back to the decision layer. At the same time, the force perception and body perception data of the joint angle position of the rotating rod (2) are also fed back to the decision layer, so that the decision layer can intelligently control the execution layer and form an adaptive adjustment.
6. The robot grasping posture adaptive adjustment device based on vision and tactile fusion according to claim 1, characterized in that: The loading structure (65) is also provided with an electric structure (651). One end of the electric structure (651) is equipped with an assembly frame (652) and the inner wall of the assembly frame (652) is provided with a splicing rail (653). A locking bolt (654) is provided on the outer side of the splicing rail (653), i.e. the outer edge of the assembly frame (652). An electromagnetic block (655) is also connected to the center inside the assembly frame (652). The energized structure (651) is perpendicular to the assembly frame (652). The assembly frame (652) is in the shape of a concave "U". The splicing rail (653) of the assembly frame (652) is set in a symmetrical orientation and is connected to the electromagnetic block (655). The electromagnetic block (655) is a raised solid shape.
7. The robot grasping posture adaptive adjustment device based on vision and tactile fusion according to claim 6, characterized in that: The energized structure (651) is also provided with a reinforcing block (6511), one end of which is connected to an insulating frame (6512). A limiting frame (6513) is mounted on the inner wall of the insulating frame (6512), and an electrical connection protrusion (6514) is included in the limiting frame (6513). The reinforcing block (6511) is a rubber product and there are two blocks arranged symmetrically at the upper and lower center positions of the insulating frame (6512). The insulating frame (6512) and the limiting frame (6513) overlap each other and cover the edge of the electrical connection protrusion (6514).
8. The robot grasping posture adaptive adjustment device based on vision and tactile fusion according to claim 1, characterized in that: The extender (3) is provided with a connecting end (31), which is integrated with the balance block (32). The upper end of the balance block (32) is provided with an overlapping layer (33), and a positioning block (34) is provided at the center of the overlapping layer (33). A sliding rod (35) is connected to the center of the positioning block (34), and an auxiliary rod (36) is connected to the edge of the overlapping layer (33). The balance block (32) is fixedly connected to one end of the intelligent control host (4) through the connecting end (31). The overlapping layer (33) of the balance block (32) is parallel to one end of the rotating rod (2). The slide rod (35) includes a small drive motor and a rotating pulley. There are four auxiliary rods (36) distributed at the corners of the overlapping layer (33) and inserted into the rotating rod (2).
9. The robot grasping posture adaptive adjustment device based on vision and tactile fusion according to claim 8, characterized in that: The auxiliary rod (36) is also provided with a locking rod (361), the upper end of the locking rod (361) is connected to a bottom block (362), the center of the upper end of the bottom block (362) is provided with a rod body (363), and the upper end of the rod body (363) is connected to an anti-detachment block (364). The locking rod (361) is provided with an external thread and forms a "T" shape with the bottom block (362). The upper end of the bottom block (362) is perpendicular to the rod body (363), and the top area of the rod body (363) is smaller than the area of the anti-detachment block (364).