CN106313045A - Learning method and system of robot - Google Patents
Learning method and system of robot Download PDFInfo
- Publication number
- CN106313045A CN106313045A CN201610849767.0A CN201610849767A CN106313045A CN 106313045 A CN106313045 A CN 106313045A CN 201610849767 A CN201610849767 A CN 201610849767A CN 106313045 A CN106313045 A CN 106313045A
- Authority
- CN
- China
- Prior art keywords
- robot
- learning
- learning data
- optimized
- data
- 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
Links
Classifications
-
- 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
-
- 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
- B25J9/1602—Programme controls characterised by the control system, structure, architecture
- B25J9/161—Hardware, e.g. neural networks, fuzzy logic, interfaces, processor
-
- 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
- B25J9/1679—Programme controls characterised by the tasks executed
Landscapes
- Engineering & Computer Science (AREA)
- Robotics (AREA)
- Mechanical Engineering (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Manipulator (AREA)
Abstract
The invention relates to the field of industrial robots, in particular to a learning method and system of a robot. The learning method comprises the following steps of: firstly, presetting a learning task which is accomplished by a master robot; generating a learning datum by the master robot according to a condition of accomplishing the learning task; guiding out the learning datum which is learned by a slave robot, and optimizing the learning datum; judging whether the learning datum is optimized by the slaver robot or not; and if the learning datum is optimized by the slaver robot, guiding out the optimized learning datum which is learned by the slaver robot, otherwise, continuously optimizing the learning datum. According to the learning method, participation of professional programming personnel is not needed, and repetitive programming is not needed, and therefore, the learning method is a more labor-saving and efficient method.
Description
Technical field
The present invention relates to industrial robot field, particularly relate to learning method and the system of a kind of robot.
Background technology
Robot is the installations automatically performing work, and it both can accept mankind commander, can run again and compile in advance
The program of row, it is also possible to according to the principle guiding principle action formulated with artificial intelligence technology.It can assist or replace the work of the mankind
Make, and be increasingly widely used among the routine work life of people.Such as produce industry, building industry, service trade or high-risk
The work of danger, or the most popular family's sweeping robot etc..As it is shown in figure 1, robot typically by mechanical part,
Transducing part, control part three parts composition, these three parts are divided into drive system, mechanical system, impression system, machine
These six subsystems of people's environmental interaction system, man-machine interactive system, control system.Robot control system is typically by one or many
Platform microcomputer is constituted to complete the control to robot.There is distinct disadvantage per capita in this kind of machine.First, complicated control
Robot must be programmed by system by the programming personnel of specialty, just can carry out actual job, once operation after programming is good
Changing, be necessary for robot reprogramming could be continued operation, this makes the burden of programming personnel be greatly increased.Its
Secondary, the robot control system platform of different vendor and standard are the most different.Such as, Saphira is that SRI is international artificial
The robot control system that intelligent centre was developed in the nineties in 20th century, supports Unix and windows platform, uses C language to open
Send out；TeamBots is an application program based on JAVA and kit set；Domestic robot system the earliest be Shenyang from
The OSMOR system of dynamicization Research Institute, it is open architecture based on Windows operating system.Visible, different factories
Programming language and the system structure of the robot of business are the most different, and this makes the work of programming personnel become difficulty the most further.
Again, simple man-machine interactive system cannot be precisely controlled robot, such as clean robot, owing to operative goals often occurs
Change, is just constantly needed to be controlled robot adjusting, but can not accomplish again accurately to adjust so that operation effectiveness is also paid no attention to
Think.Finally, traditional robot is with CPU or MCU for bottom control hardware, although possesses and necessarily stores function, but far from
Can realize the function of self-teaching, the most such robot will use the increase of time to become more " clever because of user
Bright ".Therefore, conventional machines people still haves much room for improvement.
Summary of the invention
For the problems referred to above, the present invention proposes the learning method of a kind of robot, and described robot includes main robot
With at least one from robot, described learning method includes:
Step S1, pre-sets learning tasks, and described main robot completes described learning tasks；
Step S2, described main robot generates a learning data according to the situation completing described learning tasks；
Step S3, derives described learning data, described from learning data described in robot learning, and enters described learning data
Row optimizes；
Step S4, it is judged that described whether described learning data is optimized from robot, if then performing step S5, otherwise,
Perform described step S3；
Step S5, derives the described learning data optimized, the described described learning data optimized from robot learning.
The learning method of above-mentioned robot, wherein, in step S1, described have from the artificial neural network chip of robot
Fully feel and know, learn and memory ability, so that described learning data is optimized.
The learning method of above-mentioned robot, wherein, described artificial neural network chip is realized by hardware circuit, and includes
Jumbo nonvolatile memory.
The learning method of above-mentioned robot, wherein, described learning method also includes:
Described learning data is processed so that the parameter in described learning data changes in proportion, to adapt to different size
From robot.
The learning method of above-mentioned robot, wherein, described parameter includes: ambient parameter and/or coordinate parameters and/or position
Put parameter and/or load parameter.
The learning method of above-mentioned robot, wherein, in step S1, main robot described in artificial teaching completes described study
Task, and described artificial teaching includes direct teaching and indirect teaching.
The learning method of above-mentioned robot, wherein, in step S3, the artificial described learning data derived has one to be preset
File format.
The learning method of above-mentioned robot, wherein, in step S3, the described learning data of artificial derivation, described from machine
People learns described learning data, and is optimized described learning data.
The learning method of above-mentioned robot, wherein, described main robot and described from the artificial similar humanoid robot of machine.
The learning method of above-mentioned robot, wherein, described learning tasks are deliberate action.
A kind of learning system of robot, described learning system includes:
Main robot, the learning tasks completing to preset generate learning data；
At least one, from robot, is connected with described main robot, imports described learning data from described main robot, to learn
State learning data, and described learning data is optimized；
Whether judge module, described is connected from robot with each, it is judged that described carry out excellent to described learning data from robot
Change, and the described learning data optimized is derived to described in other from robot, the described learning data that is not optimised keeps not
Become.
The learning system of above-mentioned robot, wherein, in described learning system:
Described include that artificial neural network chip, the described artificial neural network chip from robot possess sense from robot
Know, learn and memory ability, so that described learning data is optimized.
The learning system of above-mentioned robot, wherein, described artificial neural network chip is realized by hardware circuit, and includes
Jumbo nonvolatile memory.
In sum, the present invention proposes learning method and the system of a kind of robot, first pre-sets learning tasks, institute
Stating main robot and complete described learning tasks, described main robot generates a study number according to the situation completing described learning tasks
According to, then derive described learning data, described from learning data described in robot learning, and described learning data is carried out excellent
Change, it is judged that described whether described learning data is optimized from robot, if then deriving the described learning data of optimization, institute
State the described learning data optimized from robot learning, otherwise, continue described learning data is optimized, it is not necessary to specialty programming
The participation of personnel, it is not necessary to the programming of repeatability, is one more human-saving, more efficient method.
Accompanying drawing explanation
Fig. 1 is the internal structure schematic diagram of existing robot；
Fig. 2 is the Method And Principle figure of the learning method of embodiment of the present invention robot；
Fig. 3 is the structure principle chart of the learning system of embodiment of the present invention robot；
Fig. 4 is the schematic diagram of embodiment of the present invention model aircraft.
Detailed description of the invention
The present invention proposes learning method and the system of a kind of robot, and this robot can be equipped with artificial neural network core
Sheet, this artificial neural network chip can use hardware circuit to carry out simulated implementation human brain function, and distributed, parallel mode is stored up
Deposit, process information, consistent to the processing mode of information with the neuron in human brain and synapse.The present inventor's artificial neural networks chip
Can possess jumbo data storage capacities, even and if after power down data still can preserve, thus can be as brain one
Sample possesses memory function.Therefore, robot of the present invention cognition, memory and self-learning capability.When by the artificial side guided
The when that Shi Shi robot completing expection action or task, its artificial neural network chip just can in this process constantly perception and
Self study, and the mass data produced during this is preserved, permanent recording as human brain, thus robot
Just can repeat above-mentioned action or task by this memory function.If robot is big by produce in this learning process again
Amount data derive and download in the artificial neural network chip in other similar humanoid robots according to certain specific format, then
Other robot is no need for programming, just can complete above-mentioned same expection action and task without guiding.Visible, this realization
Method is without the participation of professional programmer, and need not repeat to guide, and owing to this robot of the present invention is learnt by oneself
Habit ability, such that it is able to tackle different environmental change, coordinate or change in location and load change, corresponding adjustment work journey
Sequence, makes more precisely, outstanding completes intended task.Therefore the intellectual learning method of this robot of the present invention can be significantly
Reduce cost and the difficulty of programming.
With embodiment, the present invention is further described below in conjunction with the accompanying drawings.
Embodiment one
As in figure 2 it is shown, present embodiments provide the learning method of a kind of robot, robot include main robot and at least one
From robot, this learning method may include that
Step S1, pre-sets learning tasks, and main robot completes learning tasks；
Step S2, main robot generates a learning data according to the situation completing learning tasks；
Step S3, derives learning data, from robot learning learning data, and is optimized learning data；
Step S4, it is judged that whether be optimized learning data from robot, if then performing step S5, otherwise, performs step
S3；
Step S5, derives the learning data optimized, the learning data optimized from robot learning.
Preferably, in step S1, artificial neural network chip can be included from robot, this artificial neural network core
Sheet possesses perception, study and memory ability, to be optimized learning data；It is further preferable that artificial neural network chip can
To be realized by hardware circuit, and include jumbo nonvolatile memory.
Preferably, this learning method can also include: processes learning data so that the parameter in learning data is pressed
Ratio changes, various sizes of from robot to adapt to；This preferably in the case of, main robot can even is that by necessarily
Robot that is scaling or that reduce, this robot zoomed in or out can produce after completing learning tasks and to zoom in or out
Learning data, can change the parameter in learning data in proportion after treatment, is adapted to various sizes of robot,
It is also possible that the selection of main robot reduces cost the most flexibly.
Preferably, parameter may include that ambient parameter and/or coordinate parameters and/or location parameter and/or load parameter,
These parameters can be to change with actual act or the change of task.
Preferably, in step S1, artificial teaching main robot can be used to complete learning tasks, and artificial teaching can be wrapped
Include direct teaching and indirect teaching；Directly teaching can be that operator use the action bars inserted in robot arm, by giving
Determining sequence of motion teaching movement content, robot is automatically order, position and time etc. concrete records of values in memory；Between
Connecing teaching can be to use teaching box (or claiming demonstrator) teaching, and operator completes job space rail by teaching box key manipulation
Mark point and the teaching about information such as speed thereof, then carry out the volume of user job program with operation dish to robot language order
Volume, and it is stored in training data district.
Preferably, in step S3, the artificial learning data derived has a file format preset.
Preferably, in step S3, manually derive learning data, from robot learning learning data, and to learning data
It is optimized.
Preferably, main robot and from the artificial similar humanoid robot of machine, but this is preferred situation, should not be understood as
Limitation of the present invention, dissimilar but the robot of identical working procedure can be used it will be also be appreciated that be comprised in the present invention
In.
Preferably, learning tasks are that the action completed under deliberate action, friction speed or acceleration can also be considered as difference
Action.
Embodiment two
The present embodiment additionally provides the learning system of a kind of robot, as it is shown on figure 3, this learning system may include that
Main robot 1, the learning tasks completing to preset generate learning data；
At least one from robot (the present embodiment as a example by two, i.e. from robot 21, from robot 22), with main robot even
Connect, import learning data from main robot, to learn learning data, and learning data is optimized；
Judge module 3, is connected from robot (being the present embodiment from robot 21, from robot 22) with each, it is judged that from
Robot 21, whether learning data is optimized from robot 22, and the learning data optimized is derived to other from machine
In people, the learning data being not optimised keeps constant.
Preferably, in this learning system: artificial neural network chip can be included from robot, artificial from robot
Neural network chip possesses perception, study and memory ability, to be optimized learning data.
Preferably, this artificial neural network chip can be realized by hardware circuit, and includes jumbo non-volatile deposit
Reservoir.
Embodiment three
Airplane skin, with aluminium alloy as main material, coordinates macromolecule resin material and the steel system of a part of high tensile force resistance
Become.Aircraft shell awing is glued by various suspended particles in fuel-burning gas and air and the dirt in sky cloud and mist
Attached make its shell be contaminated.So aircraft must clean on time, not so affect very big, due to aircraft altitude and area the most very
Greatly, it is common that 30 Duo Renyige groups, work simultaneously, the longest.Will be big if using industrial robot to complete above-mentioned work
Big manpower and the difficulty of reducing, as shown in Figure 3.But owing to aircraft is the hugest, and coordinate, position often change, one
As robot be difficult to carry out work by fixing program, it is necessary to just can be completed by manually guiding, so the operation of repetition is the most very
Time-consuming and increase the difficulty of operator.If using the intellectual learning method of this robot of the present invention, due to this
Bright robot perception, self study and memory function, it is only necessary to by one of them robot by the way of artificial guiding complete
The cleaning process of aircraft, the most just can derive the mass data that robot produces in above-mentioned cleaning process and import in pairs
In artificial neural network chip in the control system of other cleaning robot so that it is his robot also can complete same
Cleaning process, and these robots being equipped with artificial neural network chip can be continuous according to coordinate, the change of position
Self study, continues to optimize working procedure, and enable that aircraft cleans is cleaner.Such as, same program, because orientation or seat
Target changes the cleaning brush of cleaning robot and is probably different with the cleaning pressure of aircraft surfaces, this robot of the present invention
Just can make cleaning dynamics optimization by the pressure size of perception cleaning brush Yu aircraft surfaces by self study, so that aircraft
Clean the cleanest.Further, owing to aircraft and cleaning robot are the hugest, artificial guided robot is also to be stranded very much
Difficult, therefore the present invention can also be by the method for scaled down, by cleaning robot and aircraft being contracted with same ratio
Little, make model, then operator just can be by very easily making cleaning robot in laboratory in the way of artificial guiding
Human model completes the cleaning to model aircraft.The cleaning robot human model self study energy by its insider artificial neural networks chip
Power and mass data storage ability, derive the working procedure in above-mentioned cleaning process, then coordinate equal proportion is amplified
To original size, then the working procedure after updating imports in actual cleaning robot, thus more human-saving, convenient
Complete the programming to cleaning robot.Visible, this intellectual learning method to robot of the present invention is a kind of more efficient, more square
Just implementation method.
In sum, the present invention proposes a kind of intellectual learning method of robot, and described robot is equipped with ANN
Network chip, possesses perception, self study and mass data storage ability.By the way of artificial guided robot, it has been allowed to pre-
The task of phase, the working procedure then produced downloads to other with making other robot also be able in the robot of type
Become same expection task, it is not necessary to the participation of professional programmer, it is not necessary to the programming of repeatability, be one more human-saving, more
Efficient method.
By explanation and accompanying drawing, give detailed description of the invention and describe in detail, for a person skilled in the art, read
After reading described above, various changes and modifications will be apparent to undoubtedly.Therefore, appending claims should be regarded as and contains this
The true intention of invention and whole variations and modifications of scope.In Claims scope the scope of any and all equivalence with
Content, is all considered as still belonging to the intent and scope of the invention.
Claims (13)
1. the learning method of a robot, it is characterised in that described robot includes main robot and at least one is from machine
People, described learning method includes:
Step S1, pre-sets learning tasks, and described main robot completes described learning tasks；
Step S2, described main robot generates a learning data according to the situation completing described learning tasks；
Step S3, derives described learning data, described from learning data described in robot learning, and enters described learning data
Row optimizes；
Step S4, it is judged that described whether described learning data is optimized from robot, if then performing step S5, otherwise,
Perform described step S3；
Step S5, derives the described learning data optimized, the described described learning data optimized from robot learning.
The learning method of robot the most according to claim 1, it is characterised in that in described step S1, described from machine
The artificial neural network chip of people possesses perception, study and memory ability, to be optimized described learning data.
The learning method of robot the most according to claim 2, it is characterised in that described artificial neural network chip is by firmly
Part circuit realiration, and include jumbo nonvolatile memory.
The learning method of robot the most according to claim 1, it is characterised in that described learning method also includes:
Described learning data is processed so that the parameter in described learning data changes in proportion, to adapt to different size
From robot.
The learning method of robot the most according to claim 4, it is characterised in that described parameter includes: ambient parameter and/
Or coordinate parameters and/or location parameter and/or load parameter.
The learning method of robot the most according to claim 1, it is characterised in that in step S1, main described in artificial teaching
Robot completes described learning tasks, and described artificial teaching includes direct teaching and indirect teaching.
The learning method of robot the most according to claim 1, it is characterised in that in described step S3, artificial derivation
Described learning data has a file format preset.
The learning method of robot the most according to claim 1, it is characterised in that in described step S3, manually derives institute
State learning data, described from learning data described in robot learning, and described learning data is optimized.
The learning method of robot the most according to claim 1, it is characterised in that described main robot and described from machine
Artificial similar humanoid robot.
The learning method of robot the most according to claim 1, it is characterised in that described learning tasks are deliberate action.
The learning system of 11. 1 kinds of robots, it is characterised in that described learning system includes:
Main robot, the learning tasks completing to preset generate learning data；
At least one, from robot, is connected with described main robot, imports described learning data from described main robot, to learn
State learning data, and described learning data is optimized；
Whether judge module, described is connected from robot with each, it is judged that described carry out excellent to described learning data from robot
Change, and the described learning data optimized is derived to described in other from robot, the described learning data that is not optimised keeps not
Become.
The learning system of 12. robots according to claim 1, it is characterised in that in described learning system:
Described include that artificial neural network chip, the described artificial neural network chip from robot possess sense from robot
Know, learn and memory ability, so that described learning data is optimized.
The learning method of 13. robots according to claim 1, it is characterised in that described artificial neural network chip by
Hardware circuit realizes, and includes jumbo nonvolatile memory.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610849767.0A CN106313045A (en) | 2016-09-26 | 2016-09-26 | Learning method and system of robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610849767.0A CN106313045A (en) | 2016-09-26 | 2016-09-26 | Learning method and system of robot |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106313045A true CN106313045A (en) | 2017-01-11 |
Family
ID=57820496
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610849767.0A Pending CN106313045A (en) | 2016-09-26 | 2016-09-26 | Learning method and system of robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106313045A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113884857A (en) * | 2021-09-29 | 2022-01-04 | 上海阵量智能科技有限公司 | Chip, chip pressure testing method and device, electronic equipment and storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS60200312A (en) * | 1984-03-26 | 1985-10-09 | Hitachi Ltd | Automatic device for graphic practice or the like |
| US20050053909A1 (en) * | 2003-09-08 | 2005-03-10 | Chan Kwok Hung | Learn-and-play programming method for motorized toys and domestic appliances |
| CN101142062A (en) * | 2005-03-18 | 2008-03-12 | 株式会社安川电机 | Teaching box used for robot, customize method, and robot system using the same |
| CN203460180U (en) * | 2013-08-12 | 2014-03-05 | 刘达 | Robot capable of directly dragging and teaching |
| CN103753518A (en) * | 2014-01-10 | 2014-04-30 | 刘汝发 | Robot teaching by hands |
| CN103878772A (en) * | 2014-03-31 | 2014-06-25 | 北京工业大学 | Biomorphic wheeled robot system with simulation learning mechanism and method |
| CN104354157A (en) * | 2014-10-29 | 2015-02-18 | 南京航空航天大学 | Tire transfer robot and control method thereof |
| CN104440864A (en) * | 2014-12-04 | 2015-03-25 | 深圳先进技术研究院 | Master-slaver teleoperation industrial robot system and control method thereof |
| CN104808484A (en) * | 2015-01-20 | 2015-07-29 | 上海优爱宝机器人技术有限公司 | Robot adaptive control method |
-
2016
- 2016-09-26 CN CN201610849767.0A patent/CN106313045A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS60200312A (en) * | 1984-03-26 | 1985-10-09 | Hitachi Ltd | Automatic device for graphic practice or the like |
| US20050053909A1 (en) * | 2003-09-08 | 2005-03-10 | Chan Kwok Hung | Learn-and-play programming method for motorized toys and domestic appliances |
| CN101142062A (en) * | 2005-03-18 | 2008-03-12 | 株式会社安川电机 | Teaching box used for robot, customize method, and robot system using the same |
| CN203460180U (en) * | 2013-08-12 | 2014-03-05 | 刘达 | Robot capable of directly dragging and teaching |
| CN103753518A (en) * | 2014-01-10 | 2014-04-30 | 刘汝发 | Robot teaching by hands |
| CN103878772A (en) * | 2014-03-31 | 2014-06-25 | 北京工业大学 | Biomorphic wheeled robot system with simulation learning mechanism and method |
| CN104354157A (en) * | 2014-10-29 | 2015-02-18 | 南京航空航天大学 | Tire transfer robot and control method thereof |
| CN104440864A (en) * | 2014-12-04 | 2015-03-25 | 深圳先进技术研究院 | Master-slaver teleoperation industrial robot system and control method thereof |
| CN104808484A (en) * | 2015-01-20 | 2015-07-29 | 上海优爱宝机器人技术有限公司 | Robot adaptive control method |
Non-Patent Citations (1)
| Title |
|---|
| 陈荷娟: "《机电一体化系统设计》", 31 March 2013 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113884857A (en) * | 2021-09-29 | 2022-01-04 | 上海阵量智能科技有限公司 | Chip, chip pressure testing method and device, electronic equipment and storage medium |
| CN113884857B (en) * | 2021-09-29 | 2024-03-08 | 上海阵量智能科技有限公司 | Chip, chip pressure testing method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI707280B (en) | Action prediction system and action prediction method | |
| Asada et al. | Transfer of human skills to neural net robot controllers | |
| JP5361756B2 (en) | Robot with automatic selection of task-specific expressions for imitation learning | |
| Sahbani et al. | An overview of 3D object grasp synthesis algorithms | |
| CN103926838B (en) | Autonomous metal development cloud robot system based on cloud computing | |
| CN109483534B (en) | Object grabbing method, device and system | |
| CN111515961A (en) | Reinforcement learning reward method suitable for mobile mechanical arm | |
| Tykal et al. | Incrementally assisted kinesthetic teaching for programming by demonstration | |
| CN108406767A (en) | Robot autonomous learning method towards man-machine collaboration | |
| CN109702740B (en) | Robot compliance control method, device, equipment and storage medium | |
| CN111872934A (en) | Mechanical arm control method and system based on hidden semi-Markov model | |
| WO2020118730A1 (en) | Compliance control method and apparatus for robot, device, and storage medium | |
| CN112631128A (en) | Robot assembly skill learning method and system based on multi-mode heterogeneous information fusion | |
| Ma et al. | Can robots replace human beings?—Assessment on the developmental potential of construction robot | |
| JP2022061968A (en) | Device and method for controlling robotic device | |
| Kyrarini et al. | Human-Robot Synergy for cooperative robots | |
| Koenig et al. | Behavior-based segmentation of demonstrated tasks | |
| CN112605974A (en) | Robot complex operation skill characterization method and system | |
| Gladence et al. | A research on application of human-robot interaction using artifical intelligence | |
| Liu et al. | A mixed perception-based human-robot collaborative maintenance approach driven by augmented reality and online deep reinforcement learning | |
| CN106313045A (en) | Learning method and system of robot | |
| Grigore et al. | Predicting supportive behaviors for human-robot collaboration | |
| Kent et al. | Construction of an object manipulation database from grasp demonstrations | |
| CN110245809A (en) | A kind of intelligent optimization method and system for multirobot multitask cooperative work | |
| Cakmak et al. | Active learning with mixed query types in learning from demonstration |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170111 |