CN107336234A - A kind of reaction type self study industrial robot and method of work - Google Patents

A kind of reaction type self study industrial robot and method of work Download PDF

Info

Publication number
CN107336234A
CN107336234A CN201710441678.7A CN201710441678A CN107336234A CN 107336234 A CN107336234 A CN 107336234A CN 201710441678 A CN201710441678 A CN 201710441678A CN 107336234 A CN107336234 A CN 107336234A
Authority
CN
China
Prior art keywords
robot
self study
neutral net
pressure sensor
effector
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.)
Pending
Application number
CN201710441678.7A
Other languages
Chinese (zh)
Inventor
李泽晨
穆锐
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Saihe Intelligent Equipment (shanghai) Ltd By Share Ltd
Original Assignee
Saihe Intelligent Equipment (shanghai) Ltd By Share Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Saihe Intelligent Equipment (shanghai) Ltd By Share Ltd filed Critical Saihe Intelligent Equipment (shanghai) Ltd By Share Ltd
Priority to CN201710441678.7A priority Critical patent/CN107336234A/en
Publication of CN107336234A publication Critical patent/CN107336234A/en
Pending legal-status Critical Current

Links

Classifications

    • BPERFORMING OPERATIONS; TRANSPORTING
    • B25HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25JMANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00Programme-controlled manipulators
    • B25J9/16Programme controls
    • B25J9/1628Programme controls characterised by the control loop
    • B25J9/163Programme controls characterised by the control loop learning, adaptive, model based, rule based expert control
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B25HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25JMANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00Programme-controlled manipulators
    • B25J9/16Programme controls
    • B25J9/1602Programme controls characterised by the control system, structure, architecture
    • B25J9/161Hardware, e.g. neural networks, fuzzy logic, interfaces, processor

Abstract

The invention discloses a kind of reaction type self study industrial robot and method of work, including robot body, joint of robot, end effector of robot, pressure sensor, pedestal, vision sensor, support and robot base, the end of the robot body is provided with end effector of robot, pressure sensor is installed on end effector of robot, vision sensor is installed with the top of the support, by building neutral net, collection image and processing, control machine people performs relevant action.The reaction type self study industrial robot and method of work of the present invention, judge that whether end runs succeeded by pressure sensor data and provide learning outcome foundation for self learning neural networkses, learning track is provided for the self study of industrial robot, according to processing result image, in whole process, the trick relation of camera and robot need not be demarcated, can adaptive unknown object inter-related task, application greatly increases.

Description

A kind of reaction type self study industrial robot and method of work
Technical field
The present invention relates to industrial robot field, specially a kind of reaction type self study industrial robot and method of work.
Background technology
Conventional industrial robot realizes repeated work mainly by the method for pre-programmed come the motion of control machine people.Such as Fruit does not have predefined and programming inter-related task, and robot will be unable to make corresponding judgement and work.On this basis, mark is passed through Determine robotic vision sensor(Eye)And the end effector of robot(Hand)Between relation, i.e. trick relation, then tie The method for closing image procossing realizes that the automatic hand-eye machine people to work automatically is used widely and developed.This robot is applicable The work such as gripping, carrying in known object form, it can not automatically learn and unknown object is operated.
The content of the invention
It is an object of the invention to provide a kind of reaction type self study industrial robot and method of work, and having can be automatic The advantages of learning and being operated to unknown object, to solve the problems mentioned in the above background technology.
To achieve the above object, the present invention provides following technical scheme:A kind of reaction type self study industrial robot, including Robot body, joint of robot, end effector of robot, pressure sensor, pedestal, vision sensor, support and machine People's base, robot base are fixedly connected on pedestal, and robot base connects robot body, machine by joint of robot The end of human agent is provided with end effector of robot, is provided with pressure sensor on end effector of robot, on pedestal Support is also fixedly connected with, vision sensor is installed with the top of support.
Preferably, end performs device is installed on end effector of robot.
A kind of method of work of reaction type self study industrial robot, workflow include:
Step 1:Neutral net is built, is robot posture information, the centre coordinate of object, pressure sensor data on image As neutral net input node, whether run succeeded and be used as neutral net output node, suitable hidden node number;
Step 2:Image is gathered by vision sensor;
Step 3:Image is handled again, segmenting edge and target, obtain target dependent coordinate information such as centre coordinate;
Step 4:Then control machine people performs relevant action, if robot execution acts successfully, its pressure sensor will pass Corresponding data is sent otherwise, to perform failure to host computer;
Step 5:The information such as posture information now, processing result image, implementing result are sent to neutral net input and output, Meanwhile execution result back and coordinate information are to robot;
Step 6:Robot suitably adjusts pose according to the execution information of feedback;
Step 7:Repeat step 5 and 6, until end runs succeeded;
Step 8:Train neutral net;
Step 9:Step 2 is performed a plurality of times to step 8, until neutral net restrains, object space when obtaining pose and handling successfully, The dependency relation of pressure sensor state etc., robot self study are completed;
Step 10:After completing self study, related work is performed according to neutral net result.
Compared with prior art, the beneficial effects of the invention are as follows:The reaction type self study industrial robot and work of the present invention Make method, by pressure sensor data judge end whether run succeeded for self learning neural networkses provide learning outcome according to According to, learning track is provided for the self study of industrial robot, it is automatic to adjust according to processing result image using reaction type mode Whole robot pose, solves the problems, such as the Pose Control in automation study.In whole process, it is not necessary to demarcate camera and The trick relation of robot, dependency relation therein is as described by the hidden layer of neutral net.Can adaptive unknown object phase Pass task, application greatly increase.
Brief description of the drawings
Fig. 1 is the overall structure diagram of the present invention;
Fig. 2 is the workflow diagram of the present invention.
In figure:1 robot body, 2 joint of robot, 3 end effector of robot, 4 pressure sensors, 5 pedestals, 6 regard Feel sensor, 7 supports, 8 robot bases.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, close Chu, complete is carried out to the technical scheme in the embodiment of the present invention Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, rather than whole embodiments.It is based on Embodiment in the present invention, those of ordinary skill in the art are obtained every other under the premise of creative work is not made Embodiment, belong to the scope of protection of the invention.
Refering to Fig. 1-2, the present invention provides a kind of technical scheme:A kind of reaction type self study industrial robot, including machine Human agent 1, joint of robot 2, end effector of robot 3, pressure sensor 4, pedestal 5, vision sensor 6, support 7 and machine Device people base 8, robot base 8 is fixedly connected on pedestal 5, and robot base 8 connects robot master by joint of robot 2 Body 1, the sitting posture of control machine people, the end of robot body 1 are provided with end effector of robot 3, and robot end performs Pressure sensor 4 is installed, pressure sensor 4 is for detecting whether crawl known object, end effector of robot 3 on device 3 On performs device such as handgrip in end is installed, support 7 is also fixedly connected with pedestal 5, the top of support 7 is installed with vision Sensor 6, vision sensor 6 can be observed within the working range of end effector of robot 3.
A kind of method of work of reaction type self study industrial robot, workflow include:
Step 1:Neutral net is built, is robot posture information, the centre coordinate of object, pressure sensor 4 count on image According to as neutral net input node, whether run succeeded and be used as neutral net output node, suitable hidden node number;
Step 2:Image is gathered by vision sensor 6;
Step 3:Image is handled again, segmenting edge and target, obtain target dependent coordinate information such as centre coordinate;
Step 4:Then control machine people performs relevant action, if robot execution acts successfully, its pressure sensor 4 will pass Corresponding data is sent otherwise, to perform failure to host computer;
Step 5:The information such as posture information now, processing result image, implementing result are sent to neutral net input and output, Meanwhile execution result back and coordinate information are to robot;
Step 6:Robot suitably adjusts pose according to the execution information of feedback;
Step 7:Repeat step 5 and 6, until end runs succeeded;
Step 8:Train neutral net;
Step 9:Step 2 is performed a plurality of times to step 8, until neutral net restrains, object space when obtaining pose and handling successfully, The dependency relation of the state of pressure sensor 4 etc., robot self study are completed;
Step 10:After completing self study, related work is performed according to neutral net result.
Robot can judge unknown object shape and position by autonomous learning, and as requested, it is unknown to this Object such as is gripped, carried at the operation, and this industrial robot need not demarcate the trick relation of camera and robot, This industrial robot uses neural net method, can automatically learn and perform inter-related task, industrial robot uses reaction type Pose Control, it is not necessary to manual intervention, automatic running.
To sum up:The reaction type self study industrial robot and method of work of the present invention, is judged by the data of pressure sensor 4 Whether end runs succeeded provides learning outcome foundation for self learning neural networkses, is provided for the self study of industrial robot Learning track, using reaction type mode, according to processing result image, adjust automatically machine people's pose, solves automation study In Pose Control problem.In whole process, it is not necessary to demarcate the trick relation of camera and robot, related pass therein System is as described by the hidden layer of neutral net.Can adaptive unknown object inter-related task, application greatly increases.
More than, it is only the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, and it is any Those familiar with the art the invention discloses technical scope in, technique according to the invention scheme and its invention Design is subject to equivalent substitution or change, should all be included within the scope of the present invention.

Claims (3)

1. a kind of reaction type self study industrial robot, it is characterised in that including robot body(1), joint of robot(2)、 End effector of robot(3), pressure sensor(4), pedestal(5), vision sensor(6), support(7)And robot base (8), robot base(8)It is fixedly connected on pedestal(5)On, robot base(8)Pass through joint of robot(2)Connect robot Main body(1), robot body(1)End be provided with end effector of robot(3), end effector of robot(3)Upper peace Equipped with pressure sensor(4), pedestal(5)On be also fixedly connected with support(7), support(7)Top be installed with vision biography Sensor(6).
A kind of 2. reaction type self study industrial robot according to claim 1, it is characterised in that end effector of robot (3)On end performs device is installed.
3. a kind of method of work of reaction type self study industrial robot, it is characterised in that workflow includes:
Step 1:Neutral net is built, is robot posture information, the centre coordinate of object, pressure sensor on image(4) Whether data run succeeded as neutral net input node and are used as neutral net output node, suitable hidden node number;
Step 2:By vision sensor(6)Gather image;
Step 3:Image is handled again, segmenting edge and target, obtain target dependent coordinate information such as centre coordinate;
Step 4:Then control machine people performs relevant action, if robot execution acts successfully, its pressure sensor(4)Will Corresponding data is transmitted to host computer, otherwise, performs failure;
Step 5:The information such as posture information now, processing result image, implementing result are sent to neutral net input and output, Meanwhile execution result back and coordinate information are to robot;
Step 6:Robot suitably adjusts pose according to the execution information of feedback;
Step 7:Repeat step 5 and 6, until end runs succeeded;
Step 8:Train neutral net;
Step 9:Step 2 is performed a plurality of times to step 8, until neutral net restrains, object space when obtaining pose and handling successfully, Pressure sensor(4)The dependency relation of state etc., robot self study are completed;
Step 10:After completing self study, related work is performed according to neutral net result.
CN201710441678.7A 2017-06-13 2017-06-13 A kind of reaction type self study industrial robot and method of work Pending CN107336234A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710441678.7A CN107336234A (en) 2017-06-13 2017-06-13 A kind of reaction type self study industrial robot and method of work

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710441678.7A CN107336234A (en) 2017-06-13 2017-06-13 A kind of reaction type self study industrial robot and method of work

Publications (1)

Publication Number Publication Date
CN107336234A true CN107336234A (en) 2017-11-10

Family

ID=60221454

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710441678.7A Pending CN107336234A (en) 2017-06-13 2017-06-13 A kind of reaction type self study industrial robot and method of work

Country Status (1)

Country Link
CN (1) CN107336234A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109040688A (en) * 2018-08-23 2018-12-18 顺德职业技术学院 The method and system that the industrial robot operation video of a kind of pair of acquisition is stored
CN109079786A (en) * 2018-08-17 2018-12-25 上海非夕机器人科技有限公司 Mechanical arm grabs self-learning method and equipment
CN110302981A (en) * 2019-06-17 2019-10-08 华侨大学 A kind of solid waste sorts online grasping means and system
CN111003380A (en) * 2019-12-25 2020-04-14 深圳蓝胖子机器人有限公司 Method, system and equipment for intelligently recycling garbage

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030233170A1 (en) * 2002-03-15 2003-12-18 Shinya Ohtani Memory system, memory method, and robotic apparatus
CN101081512A (en) * 2006-05-29 2007-12-05 发那科株式会社 Work-piece picking device and method
CN104786226A (en) * 2015-03-26 2015-07-22 华南理工大学 Posture and moving track positioning system and method of robot grabbing online workpiece
CN106097322A (en) * 2016-06-03 2016-11-09 江苏大学 A kind of vision system calibration method based on neutral net
CN106272427A (en) * 2016-09-12 2017-01-04 安徽理工大学 A kind of industrial robot intelligence picking up system
CN106393102A (en) * 2015-07-31 2017-02-15 发那科株式会社 Machine learning device, robot system, and machine learning method

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030233170A1 (en) * 2002-03-15 2003-12-18 Shinya Ohtani Memory system, memory method, and robotic apparatus
CN101081512A (en) * 2006-05-29 2007-12-05 发那科株式会社 Work-piece picking device and method
CN104786226A (en) * 2015-03-26 2015-07-22 华南理工大学 Posture and moving track positioning system and method of robot grabbing online workpiece
CN106393102A (en) * 2015-07-31 2017-02-15 发那科株式会社 Machine learning device, robot system, and machine learning method
CN106097322A (en) * 2016-06-03 2016-11-09 江苏大学 A kind of vision system calibration method based on neutral net
CN106272427A (en) * 2016-09-12 2017-01-04 安徽理工大学 A kind of industrial robot intelligence picking up system

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109079786A (en) * 2018-08-17 2018-12-25 上海非夕机器人科技有限公司 Mechanical arm grabs self-learning method and equipment
CN109040688A (en) * 2018-08-23 2018-12-18 顺德职业技术学院 The method and system that the industrial robot operation video of a kind of pair of acquisition is stored
CN109040688B (en) * 2018-08-23 2020-09-25 顺德职业技术学院 Method and system for storing acquired industrial robot operation video
CN110302981A (en) * 2019-06-17 2019-10-08 华侨大学 A kind of solid waste sorts online grasping means and system
CN111003380A (en) * 2019-12-25 2020-04-14 深圳蓝胖子机器人有限公司 Method, system and equipment for intelligently recycling garbage

Similar Documents

Publication Publication Date Title
Sanchez et al. Robotic manipulation and sensing of deformable objects in domestic and industrial applications: a survey
Pathak et al. Zero-shot visual imitation
Ciocarlie et al. Towards reliable grasping and manipulation in household environments
US9792546B2 (en) Hierarchical robotic controller apparatus and methods
El Zaatari et al. Cobot programming for collaborative industrial tasks: An overview
JP6685999B2 (en) System and method for flexible human-machine collaboration
DE102013203381B4 (en) Method and system for training an robot using a responsible demonstration supported by people
Kalakrishnan et al. Learning force control policies for compliant manipulation
Yang et al. A learning framework of adaptive manipulative skills from human to robot
CN102814814B (en) Kinect-based man-machine interaction method for two-arm robot
Zollner et al. Programming by demonstration: Dual-arm manipulation tasks for humanoid robots
Popović et al. A strategy for grasping unknown objects based on co-planarity and colour information
Leeper et al. Strategies for human-in-the-loop robotic grasping
Tolani et al. Real-time inverse kinematics of the human arm
CN104858876B (en) Visual debugging of robotic tasks
CN103271784B (en) Man-machine interactive manipulator control system and method based on binocular vision
JP2017080886A (en) Methods and computer-program products for evaluating grasp pattern, and robot incorporating the same
CN103112007B (en) Based on the man-machine interaction method of hybrid sensor
CN106737662B (en) Robot system
US8965580B2 (en) Training and operating industrial robots
Ekvall et al. Learning and evaluation of the approach vector for automatic grasp generation and planning
Hersch et al. Dynamical system modulation for robot learning via kinesthetic demonstrations
Kohlbrecher et al. Human‐robot teaming for rescue missions: Team ViGIR's approach to the 2013 DARPA Robotics Challenge Trials
CN106945043B (en) Multi-arm cooperative control system of master-slave teleoperation surgical robot
KR20130083099A (en) Robot and method to recognize and handle exceptional situations

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
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20171110

WD01 Invention patent application deemed withdrawn after publication