CN109551507B - Software manipulator based on machine learning - Google Patents
Software manipulator based on machine learning Download PDFInfo
- Publication number
- CN109551507B CN109551507B CN201910037142.8A CN201910037142A CN109551507B CN 109551507 B CN109551507 B CN 109551507B CN 201910037142 A CN201910037142 A CN 201910037142A CN 109551507 B CN109551507 B CN 109551507B
- Authority
- CN
- China
- Prior art keywords
- soft
- gear
- finger
- optical fiber
- mechanical finger
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J15/00—Gripping heads and other end effectors
- B25J15/08—Gripping heads and other end effectors having finger members
- B25J15/12—Gripping heads and other end effectors having finger members with flexible finger members
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
- B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
- B25J9/00—Programme-controlled manipulators
- B25J9/16—Programme controls
- B25J9/1628—Programme controls characterised by the control loop
- B25J9/163—Programme controls characterised by the control loop learning, adaptive, model based, rule based expert control
Abstract
The embodiment of the application provides a soft mechanical arm based on machine learning, wherein the soft mechanical finger comprises a hollow mechanical finger and a hollow soft finger sleeve, the inner wall of the finger sleeve is coated with a reflective material, the upper end of the finger sleeve is provided with an opening, the finger sleeve is sleeved at the tail end of the mechanical finger through the opening, an optical fiber is arranged in the hollow structure of the mechanical finger and extends into the soft finger sleeve, the optical fiber senses the deformation generated by the reflective material on the inner wall of the finger sleeve, an optical fiber sending end transmits an image to an optical fiber receiving end through an optical fiber array, the optical fiber receiving end is connected with a control system, the control system receives the image, judges a captured biological sample through machine learning and controls an executing mechanism to perform capturing work through a driving mechanism, the invention can enable the mechanical arm to autonomously judge the type and the weight of the captured biological sample through a, the integrity of the biological sample is ensured.
Description
Technical Field
The application relates to the technical field of manipulators, in particular to a soft manipulator based on machine learning.
Background
The function and form of the conventional rigid manipulator have been widely used in the production of human society. These rigid manipulators are mainly made of metal materials and are often responsible for structural, repetitive and other work in the industrial field. Modern manipulators, however, have evolved towards high positioning accuracy, high flexibility and high human-machine interaction. In recent years, soft manipulators designed and manufactured by using soft materials attract wide attention of scholars and institutions at home and abroad and are continuously researched, so that a solution idea and a solution direction are provided for solving the problems of poor environmental interactivity, poor complex environmental adaptability and the like of the rigid manipulator essentially. Researchers at the university of chicago and the university of cornell developed a "coffee pack" soft hand that grips objects by occlusion technology, and the flexible manipulator enabled the gripping of a variety of complex shaped objects through a change in stiffness. DEIMEL et al, at berlin university, developed a pneumatic driver-based soft body manipulator RBO Hand that trained the soft body to grab flexibly by enhancing the learning algorithm. The development of marine resources has gradually become a target of many countries, and the manipulator is an important means for collecting marine resources and marine organism samples for the development and utilization of marine resources. The rigid mechanical arm is easy to cause damage when collecting marine biological samples, and the integrity of the biological samples cannot be obtained.
Disclosure of Invention
The application provides a software manipulator based on machine learning can make the manipulator independently judge the kind and the weight of snatching the biological sample through the mode of machine learning to give suitable dynamics and snatch, guaranteed the integrality of biological sample.
The utility model provides a software manipulator based on machine learning, including actuating mechanism, actuating mechanism and control system, wherein, actuating mechanism is software mechanical finger, software mechanical finger includes hollow mechanical finger and hollow software dactylotheca, the inner wall of dactylotheca scribbles reflecting material, the upper end of dactylotheca is provided with the opening, through the opening with the dactylotheca cover is at the end of mechanical finger, be provided with optic fibre in the hollow structure of mechanical finger, optic fibre extends to in the software dactylotheca, the optic fibre perception the deformation that reflecting material of dactylotheca inner wall produced, optic fibre sending terminal passes through the fiber array and conveys the image to the optic fibre receiving terminal, the control system is connected to the optic fibre receiving terminal, control system receives the image, judges the biological sample who snatchs through machine learning, and the actuating mechanism is controlled by the driving mechanism to carry out grabbing work.
Further, the driving mechanism comprises a motor, a power gear, a first gear, a second gear, a first power arm, a second power arm and a guide rod, the motor is connected with the power gear, the power gear is meshed with the second gear, the second gear is meshed with the first gear, one end of the first power arm is fixed at the center of the first gear and is integrated with the first gear, the other end of the first power arm extends out of the first gear, a through hole is formed in the other end of the first power arm, the guide rod penetrates through the through hole in the other end of the first power arm and the through hole in the mechanical finger, so that the first power arm is linked with the mechanical finger, one end of the second power arm is fixed at the center of the second gear and is integrated with the second gear, and the guide rod penetrates through the through hole in the other end of the second power arm and the through hole in the mechanical finger, so that the second power arm is linked with the mechanical finger.
As a preferred technical scheme, the soft finger stall is a silica gel finger stall.
As a preferred technical scheme, the light reflecting material is a tin foil material.
As a preferred technical solution, there are two pairs of optical fibers inside each mechanical finger, where one pair of optical fibers is used for sending a stress variation image, and the other pair of optical fibers is used for receiving the stress variation image.
The software manipulator based on machine learning that this application embodiment provided can make the manipulator independently judge the kind and the weight of the biological sample of snatching through the mode of machine learning to give suitable dynamics and snatch, guaranteed the integrality of biological sample.
Drawings
In order to more clearly explain the technical solution of the present application, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious to those skilled in the art that other drawings can be obtained according to the drawings without any creative effort.
Fig. 1 is a schematic structural diagram of a soft manipulator according to an embodiment of the present disclosure;
FIG. 2 is a front view of a soft body robot according to an embodiment of the present disclosure;
fig. 3 is a schematic structural view of a soft finger cot provided in the embodiment of the present application;
fig. 4 is a schematic structural diagram of a mechanical finger according to an embodiment of the present disclosure;
FIG. 5 is a schematic view of a gear structure provided in an embodiment of the present application;
FIG. 6 is a schematic structural diagram of a power arm and a guide rod according to an embodiment of the present disclosure;
fig. 7 is an exploded view provided in an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Examples
Referring to fig. 1, a schematic structural diagram of a soft mechanical hand shown in the present embodiment, a front view of the soft mechanical hand shown in fig. 2, and an exploded view of fig. 7, the soft mechanical hand based on machine learning includes an executing mechanism, a driving mechanism, and a control system, wherein the executing mechanism is a soft mechanical finger, the soft mechanical finger includes a hollow mechanical finger 1 and a soft finger sleeve 2, the soft finger sleeve 2 is a hollow cylindrical structure, please refer to fig. 3, (of course, the soft finger sleeve 2 may be other structures, such as a hollow rectangular structure, etc.), an inner wall of the soft finger sleeve 2 is coated with a reflective material, in the present embodiment, a tin foil is selected as the reflective material to be coated on an inner wall of the soft finger sleeve 2, an opening is disposed at an upper end of the soft finger sleeve 2, the soft finger sleeve 2 is sleeved on a terminal end of the mechanical finger 1 through the opening, and the mechanical finger 1 is also, in this embodiment, the mechanical finger 1 is a hollow rectangular structure (of course, the mechanical finger 1 may be other structures, such as a hollow cylindrical structure, etc.), the end of the mechanical finger 1 is open, the optical fiber 3 is disposed inside the hollow structure of the mechanical finger 1, specifically, refer to fig. 4, in this embodiment, four optical fibers 3 are embedded inside each hollow mechanical finger 1, the four optical fibers are two pairs of transmitting and receiving ends, the optical fibers extend into the soft finger stall, when the soft mechanical finger grabs a biological sample, the soft finger stall 2 deforms, the optical fibers 3 can sense the stress change of the reflective material on the inner wall of the soft finger stall 2, and the optical fiber transmitting end transmits an image (image of the stress change) to the optical fiber receiving end through the optical fiber array. The optical fiber receiving end is connected to a control system, the control system receives the stress change image, judges the captured biological sample through machine learning, and controls the actuating mechanism to capture through the driving mechanism.
The number and the position relation of the mechanical fingers can be adjusted according to actual needs, and in the embodiment of the application, the structure of three mechanical fingers on the left side and two mechanical fingers on the right side is adopted.
In the prior art, a pressure sensor is mostly adopted to sense stress change, the pressure sensor is a device or a device which can sense pressure signals and can convert the pressure signals into usable output electric signals according to a certain rule, the sensors are all linear sensors, the accuracy is poor, if the very good sensor is adopted, the price is very high, the embodiment of the application adopts the mode that the inner wall of the soft finger cot 2 is coated with a reflective material, the hollow inner part of the soft finger cot 2 is provided with an optical fiber to transmit images, and then the biological sample is identified through a machine learning mode, so that the method can process the linear condition and the nonlinear condition.
In the embodiment, four optical fibers are embedded inside each hollow mechanical finger, namely two groups of transceiving signal ends, so as to realize transmission of differential signals, namely a transmitting signal end TX +, a TX-and a receiving signal end RX +, an RX-, and the mode of the differential signal transmission has strong noise resistance and is beneficial to ensuring the integrity of signal transmission.
In the embodiment of the application, sea cucumbers and scallops are used as the biological samples to be grabbed, the grabbed biological samples selected here have completely different forms and structures, of course, more kinds of other biological samples can be selected as the grabbed samples, 500 training samples are selected from the sea cucumbers, 500 training samples are selected from the scallops, each training sample is randomly selected from weight and shape, during training, the training samples are grabbed in a proper mode and strength each time, stress change images of the 1000 training samples transmitted by the optical fiber array are stored, the characteristics of the training samples are recorded, in the embodiment of the application, two characteristics are selected, one is a kind, the other is weight, namely, the type and the weight of the grabbed biological sample are recorded, the amount of the weight is recorded, and the grabbing strength is large at the moment (namely, the grabbing angle of a manipulator is controlled by a motor), and repeating the experiment for 1000 times, wherein the images of 1000 training samples are used as a training data set of the software manipulator, each training sample corresponds to one type, one weight and the manipulator grabbing strength, and the manipulator can automatically judge the type and the weight of the biological sample only by receiving the images sent by the optical fiber array when grabbing scallops or sea cucumbers through machine learning of the control system, so that the biological sample can be grabbed in a proper grabbing strength and mode.
In this embodiment, the soft finger cot 2 is a silica gel finger cot, which has a certain thickness and elasticity, and can include the integrity of the grasped biological sample.
The driving mechanism (shown in fig. 5 and 6) of the soft manipulator comprises a motor 4, a power gear 5, a first gear 6, a second gear 7, a first power arm 8, a second power arm 9, a guide rod 10 and framework materials of the whole manipulator: an elastic beam (not shown in the drawing), and a fixed bracket 11, wherein the fixed bracket 11 is used for fixing the positions of the power gear 5, the first gear 6, the second gear 7, the first power arm 8, the second power arm 9, and other components and fixing other peripheral structures, the motor 4 is a stepping motor, the motor 4 is arranged on one side of the fixed bracket 11 and is directly connected with the power gear 5, the power gear 5 is engaged with the second gear 7, the second gear 7 is engaged with the first gear 6, (alternatively, the power gear 5 is engaged with the first gear 6, the first gear 6 is engaged with the second gear 7), one end of the first power arm 8 is fixed in the center of the first gear 6 and is integrated with the first gear 6, the other end of the first power arm 8 extends out of the first gear 6, the other end of the first power arm 8 is provided with a through hole, the middle part and the upper end of each mechanical finger 1 are also provided with a through hole, the guide rod 10 passes through a through hole at the other end of the first power arm 8 and a through hole on the mechanical finger, so that the first power arm 8 is linked with the mechanical finger 1, one end of the second power arm 9 is fixed at the center of the second gear 7 and is integrated with the second gear 7, the other end of the second power arm 9 extends out of the second gear 7, the other end of the second power arm 9 is provided with a through hole, the middle part and the upper end of each mechanical finger 1 are also provided with a through hole, the guide rod 10 passes through the through hole at the other end of the second power arm 9 and the through hole on the mechanical finger, so that the second power arm 9 is linked with the mechanical finger 1. In the figure, the first gear 6 is a left-side gear, the second gear 7 is a right-side gear, the first power arm 8 is a left-side power arm, and the second power arm 9 is a right-side power arm.
When the object is grabbed by the software manipulator, the software dactylotheca 2 will take place deformation, reflecting material on the inner wall of software dactylotheca 2 deforms thereupon, extend to the deformation that the reflecting material of the inside 3 perceptions of optic fibre of software dactylotheca 2 arrives the dactylotheca inner wall produced, the optic fibre sending terminal passes through the fiber array and conveys the stress variation image to the optic fibre receiving terminal, optic fibre receiving terminal connection control system, control system judges the image received, can independently judge biological sample's kind and weight, thereby snatch this biological sample with suitable dynamics and mode of snatching.
To sum up, the embodiment of the present application provides a soft mechanical arm based on machine learning, wherein the soft mechanical finger comprises a hollow mechanical finger and a hollow soft finger sleeve, the inner wall of the finger sleeve is coated with a reflective material, the upper end of the finger sleeve is provided with an opening, the finger sleeve is sleeved at the end of the mechanical finger through the opening, an optical fiber is arranged inside the hollow structure of the mechanical finger, the optical fiber extends into the soft finger sleeve, the optical fiber senses the deformation generated by the reflective material on the inner wall of the finger sleeve, the optical fiber transmitting end transmits an image to the optical fiber receiving end through an optical fiber array, the optical fiber receiving end is connected with a control system, the control system receives the image, determines a captured biological sample through machine learning, and controls an actuating mechanism to capture work through a driving mechanism, the invention can enable the mechanical arm to autonomously determine the type and, thereby giving appropriate strength to grasp and ensuring the integrity of the biological sample.
It should be noted that when the embodiments of the present application refer to the ordinal numbers "first", "second", "third", or "fourth", etc., it should be understood that the terms are used for distinguishing them from each other only, unless they really mean that the order is expressed according to the context. The terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a biological subject or device that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, biological subject or device. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, biological sample, or apparatus that comprises the element. It will be understood that the terms "central," "longitudinal," "lateral," "length," "width," "thickness," "upper end," "distal," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," and the like are used in an orientation or positional relationship indicated in the drawings for convenience in describing and simplifying the description, and do not indicate or imply that the device or element so referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be construed as limiting the invention.
The above description is merely exemplary of the present application and is presented to enable those skilled in the art to understand and practice the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
It is to be understood that the present application is not limited to what has been described above, and that various modifications and changes may be made without departing from the scope thereof. The scope of the application is limited only by the appended claims.
Claims (5)
1. A soft mechanical arm based on machine learning comprises an executing mechanism, a driving mechanism and a control system, wherein the executing mechanism is a soft mechanical finger, and the soft mechanical finger is characterized in that the soft mechanical finger comprises a hollow mechanical finger and a hollow soft fingerstall, the inner wall of the fingerstall is coated with a reflective material, the upper end of the fingerstall is provided with an opening, the fingerstall is sleeved at the tail end of the mechanical finger through the opening, an optical fiber is arranged in the hollow structure of the mechanical finger and extends into the soft fingerstall, the optical fiber senses the deformation generated by the reflective material on the inner wall of the fingerstall, an optical fiber sending end transmits an image to an optical fiber receiving end through an optical fiber array, the optical fiber receiving end is connected with the control system, the control system receives the image and judges a captured biological sample through machine learning, and the actuating mechanism is controlled by the driving mechanism to carry out grabbing work.
2. The soft manipulator according to claim 1, wherein the driving mechanism comprises a motor, a power gear, a first gear, a second gear, a first power arm, a second power arm, and a guide rod, the motor is connected to the power gear, the power gear is engaged with the second gear, the second gear is engaged with the first gear, one end of the first power arm is fixed to the center of the first gear and is integrated with the first gear, the other end of the first power arm extends out of the first gear, the other end of the first power arm is provided with a through hole, the guide rod passes through the through hole at the other end of the first power arm and the through hole on the mechanical finger, so that the first power arm is linked with the mechanical finger, one end of the second power arm is fixed to the center of the second gear, the guide rod penetrates through a through hole at the other end of the second power arm and a through hole in the mechanical finger, so that the second power arm is linked with the mechanical finger.
3. The soft manipulator of claim 1, wherein the soft finger cot is a silicone finger cot.
4. The soft manipulator of claim 1, wherein the light reflective material is a tin foil material.
5. The soft manipulator of claim 1, wherein each manipulator finger has two pairs of optical fibers therein, one pair of optical fibers is used for transmitting the stress variation image, and the other pair of optical fibers is used for receiving the stress variation image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910037142.8A CN109551507B (en) | 2019-01-15 | 2019-01-15 | Software manipulator based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910037142.8A CN109551507B (en) | 2019-01-15 | 2019-01-15 | Software manipulator based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109551507A CN109551507A (en) | 2019-04-02 |
| CN109551507B true CN109551507B (en) | 2020-12-29 |
Family
ID=65873059
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910037142.8A Active CN109551507B (en) | 2019-01-15 | 2019-01-15 | Software manipulator based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109551507B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110587571A (en) * | 2019-10-18 | 2019-12-20 | 南京蜘蛛侠智能机器人有限公司 | Mechanical arm trolley |
| CN110702631A (en) * | 2019-11-20 | 2020-01-17 | 江西中医药大学 | NIR-based online traditional Chinese medicine water content measuring method and system |
| CN113146660A (en) * | 2021-04-08 | 2021-07-23 | 清华大学深圳国际研究生院 | Mechanical claw for tactile perception by depth vision |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9744055B2 (en) * | 2014-04-10 | 2017-08-29 | The University Of Akron | Antagonistically actuated shape memory alloy manipulator |
| CN104552324A (en) * | 2015-01-05 | 2015-04-29 | 张广山 | Mechanical arm pliers |
| JP2018027581A (en) * | 2016-08-17 | 2018-02-22 | 株式会社安川電機 | Picking system |
| CN106891341A (en) * | 2016-12-06 | 2017-06-27 | 北京臻迪科技股份有限公司 | A kind of underwater robot and catching method |
| CN106962304A (en) * | 2017-05-02 | 2017-07-21 | 山东西霞口海珍品股份有限公司 | A kind of precious marine product under water fishes device |
| CN107225587B (en) * | 2017-06-27 | 2021-02-19 | 哈尔滨工程大学 | Shape self-adaptive manipulator structure for nondestructive fishing of benthos |
| CN107745389B (en) * | 2017-09-08 | 2020-09-11 | 燕山大学 | Finger operating mechanism for rigid-flexible-soft force measuring robot |
| CN207643152U (en) * | 2017-12-25 | 2018-07-24 | 珠海市俊凯机械科技有限公司 | Manipulator with micro pressure perceptional function |
| CN108542469B (en) * | 2018-02-14 | 2020-04-28 | 天津大学 | Image feedback-based six-dimensional force sensor, clamping probe and clamping instrument |
| CN207953923U (en) * | 2018-03-20 | 2018-10-12 | 吉林铁道职业技术学院 | A kind of automatic mechanical hand |
| CN108312172A (en) * | 2018-04-16 | 2018-07-24 | 苏州迪天机器人自动化有限公司 | A kind of flexible manipulator |
-
2019
- 2019-01-15 CN CN201910037142.8A patent/CN109551507B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN109551507A (en) | 2019-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109551507B (en) | Software manipulator based on machine learning | |
| CN109176586B (en) | Self-adaptive flexible paw based on torsion spring and robot | |
| EP2919948B1 (en) | Hand controller device | |
| CN105150225A (en) | Pinching-holding composite and adaptive robot finger device with rod and wheels connected in parallel | |
| US20090025502A1 (en) | Manipulator and robot | |
| CN108121450B (en) | Magneto-rheological force feedback data glove based on transmission of wire tube and steel wire rope | |
| Kim et al. | RoboRay hand: A highly backdrivable robotic hand with sensorless contact force measurements | |
| CN106346500A (en) | Underactuated humanoid dexterous robotic hand device | |
| CN107965634B (en) | Flexible pipeline crawling robot based on artificial muscles | |
| CN105150231A (en) | Pinching-holding composite and adaptive robot finger device with seven connecting rods connected in parallel | |
| JP2016024707A (en) | Tactile sense presentation device and inner force sense presentation system | |
| CN109834719A (en) | A method of measurement humanoid dexterous finger finger tip stress | |
| CN111872962A (en) | Full-drive humanoid hand device with multipoint array pressure sensor | |
| CN107972022B (en) | Humanoid manipulator framework | |
| CN109807915A (en) | A kind of under-actuated bionic manipulator based on the stiffness variable for becoming born of the same parents' principle | |
| CN109760087B (en) | Intelligent multifunctional two-finger bionic smart hand | |
| CN202825603U (en) | Multifunctional grabbing device | |
| CN114750130B (en) | Flexible driving palm exoskeleton robot for nerve rehabilitation | |
| Latif et al. | Design and development of muscle and flex sensor controlled robotic hand for disabled persons | |
| CN211044788U (en) | Demonstration system | |
| US20230033779A1 (en) | Robotic hand sensitive to forces in an aquatic environment | |
| CN212287680U (en) | Sliding block flexible piece flat clamping indirect self-adaptive robot hand device | |
| CN112809721B (en) | Flexible cable traction under-actuated humanoid hand based on non-uniform elastomer configuration | |
| Hikiji et al. | Hand-shaped force interface for human-cooperative mobile robot | |
| RU192179U1 (en) | Bionic hand |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |