CN109993308A - Learning system and method, shared platform and method, medium are shared based on cloud platform - Google Patents
Learning system and method, shared platform and method, medium are shared based on cloud platform Download PDFInfo
- Publication number
- CN109993308A CN109993308A CN201910248301.9A CN201910248301A CN109993308A CN 109993308 A CN109993308 A CN 109993308A CN 201910248301 A CN201910248301 A CN 201910248301A CN 109993308 A CN109993308 A CN 109993308A
- Authority
- CN
- China
- Prior art keywords
- model
- privately owned
- terminal
- cloud
- fusion calculation
- 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
- B25J9/1656—Programme controls characterised by programming, planning systems for manipulators
- B25J9/1664—Programme controls characterised by programming, planning systems for manipulators characterised by motion, path, trajectory planning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/02—Computing arrangements based on specific mathematical models using fuzzy logic
- G06N7/06—Simulation on general purpose computers
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
Abstract
The present invention provides a kind of robots based on cloud platform to share learning system, it includes that privately owned model generates terminal and cloud fusion calculation shared platform, privately owned model generates the environmental characteristic information that terminal is used for the privately owned model locally generated and is used to generate privately owned model and collects and uploads cloud fusion calculation shared platform, cloud fusion calculation shared platform includes Model Fusion computing module, Model Fusion computing module is used for the privately owned model that will be uploaded, in conjunction with other robot terminal, privately owned model generates terminal and/or other privately owned models generate the Share Model stored in the environmental characteristic information and cloud fusion calculation shared platform that terminal uploads and carry out fusion calculation, generate new Share Model, new Share Model is used for for other robot terminal, privately owned model generates terminal and/or other privately owned models generate terminal downloads and/or study.
Description
Technical field
Share learning system and method the present invention relates to a kind of robot, it is especially a kind of for warehouse logistics based on cloud
Learning system and method are shared by the robot of platform.
Background technique
Robot navigation refers to that given one target point of robot, robot can without impinging on obstacle reach target
Point, while path should be made as short as possible again.Learn to be current in the decision model for carrying out path planning based on intensified learning
More advanced method.Intensified learning is that intelligent body (Agent) is learnt in a manner of " trial and error ", is obtained by interacting with environment
Behavior is instructed in the award obtained, and target is that intelligent body is made to obtain maximum award, and intensified learning is different from connectionism study
Supervised learning is mainly manifested on enhanced signal, in intensified learning by environment provide enhanced signal be to generation movement it is good
It is bad to make a kind of evaluation (usually invariant signal), rather than tell reinforcement learning system RLS (reinforcement learning
System) how to go to generate correct movement.Due to external environment provide information it is seldom, RLS must lean on the experience of itself into
Row study.In this way, RLS obtains knowledge in the environment of action-critic, improves action scheme to adapt to environment.So
And this mode still remains some disadvantages, wherein just including being limited by that the setting of training environment, training time be longer, Wu Fali
The experience etc. learnt with former or other robot.
Robot cloud technology of sharing is some technologies solved the problems, such as in robot field in conjunction with cloud computing technology.It utilizes
It is long that robot cloud technology of sharing can efficiently solve the training time in the training in robot navigation's decision model, can not carry out
The problem of experience merges.Chinese patent CN108801269A proposes a kind of indoor cloud Algorithms of Robots Navigation System, but the system
There is no method is proposed in terms of specific path planning, method is only proposed in terms of Orientation on map, can not solve to instruct
Practice the problem of time is long and experience merges.There is presently no navigated about robot by cloud technology of sharing intensified learning
The invention of decision model learning system.
Summary of the invention
The technical problem to be solved by the present invention is to robot navigation be limited by the setting of training environment, training time it is longer,
The experience of former or other robot study can not be utilized.
In order to solve the above technical problems, the present invention provides a kind of shared learning system of the robot based on cloud platform, packet
It includes privately owned model and generates terminal and cloud fusion calculation shared platform, the privately owned model generates the private that terminal is used to locally to generate
There are model and the environmental characteristic information collected for generating the privately owned model to upload the cloud fusion calculation shared platform, institute
Stating cloud fusion calculation shared platform includes Model Fusion computing module, and the Model Fusion computing module is described for that will upload
Privately owned model generates terminal in conjunction with other robot terminal, the privately owned model and/or other privately owned models generates terminal and upload
Environmental characteristic information and the cloud fusion calculation shared platform on the Share Model that stores carry out fusion calculation, generate new be total to
Model is enjoyed, the new Share Model is used for for other robot terminal, the privately owned model generates terminal and/or other are privately owned
Model generates terminal downloads and/or study.
Preferred embodiment according to the present invention, it includes characteristic collection module, environmental simulation mould that the privately owned model, which generates terminal,
Block, reinforcing and transfer learning module, the characteristic collection module are used for collecting environmental characteristic information, the environmental simulation module
In utilizing the environmental characteristic information build environment model, the reinforcing and transfer learning module include intensified learning unit, institute
It states intensified learning unit and carries out intensified learning for being inputted after the environmental characteristic information on the environmental model, described in output
Privately owned model, the privately owned model include that the environmental model and the privately owned model generate terminal for environmental model life
At navigation strategy.
Preferred embodiment according to the present invention, the reinforcing and transfer learning module further include migration computing unit, described to move
It moves computing unit to be used for after by the Share Model of the environmental characteristic information input to downloading, carries out migration calculating, output is new
Environmental characteristic information, intensified learning is carried out for the intensified learning unit, to export new privately owned model, the privately owned model
It generates terminal and the new privately owned model is also uploaded to the cloud fusion calculation shared platform with further progress fusion calculation.
Preferred embodiment according to the present invention, the privately owned model generate terminal and use intelligent robot terminal, computer terminal
Or other intelligent terminals.
Preferred embodiment according to the present invention, the feature collection module is using described in camera and/or laser radar acquisition
The characteristic information that environmental characteristic information or the emulation for directly receiving the input of other devices construct is as the environmental characteristic information.
Preferred embodiment according to the present invention, the environmental simulation module construct simulated environment using gazebo simulation software.
Preferred embodiment according to the present invention, the Model Fusion computing module include model normalizing unit, fusion calculation list
Member, the model normalizing unit are used to the privately owned model generate the environmental characteristic information input described in terminal upload
The Share Model stored on privately owned model and the cloud fusion calculation shared platform obtains output result further to described
Output result is normalized, and carries out Confidence evaluation to the output result with comentropy, and be based on the confidence
The evaluation score of degree evaluation is weighted summation and obtains label score, and the fusion calculation unit is generated using the privately owned model
The label score of the environmental characteristic information and model normalizing unit output that terminal uploads carries out fusion calculation and obtains
To the new Share Model.
Preferred embodiment according to the present invention, it further includes first communication module, the first data that the privately owned model, which generates terminal,
Receiving module, data uploading module, the cloud fusion calculation shared platform further includes second communication module, the first communication mould
Block is communicated for obtaining the cloud fusion calculation shared platform network address with the second of the cloud fusion calculation shared platform
Module establishes communication, and the data uploading module is used for logical in the first communication module and the cloud fusion calculation shared platform
The privately owned model is uploaded the cloud fusion calculation shared platform after establishing by letter, first data reception module for
The Share Model in the cloud fusion calculation shared platform is downloaded after the cloud fusion calculation shared platform connection setup
Terminal is generated to the privately owned model.
Preferred embodiment according to the present invention, the cloud fusion calculation shared platform further include the second data reception module, number
According to download module, model memory module, second data reception module is used to communicate with the privately owned model generation terminal
The privately owned model is generated into the privately owned model that terminal uploads after foundation and receives the cloud fusion calculation shared platform, institute
Data download module is stated for generating after terminal connection setup with the privately owned model by the cloud fusion calculation shared platform
On the Share Model download to the privately owned model and generate terminal, the model memory module is for storing the cloud fusion
Calculate the Share Model in shared platform.
Preferred embodiment according to the present invention, the cloud fusion calculation shared platform includes Model Fusion computing module, described
Model Fusion computing module is used for robot terminal and/or the privately owned model of other intelligent terminals upload, environmental characteristic information
Fusion calculation is carried out with the Share Model in the cloud fusion calculation shared platform, generates new Share Model, described new is total to
Enjoy model for supply the robot terminal and/or other intelligent terminals and/or other robot terminal downloads and/or study.
Preferred embodiment according to the present invention, the Model Fusion computing module include model normalizing unit, fusion calculation list
Member, the environment that the model normalizing unit is used to upload the robot terminal and/or other described intelligent terminals are special
Reference breath is input to the privately owned model that the robot terminal and/or other described intelligent terminals upload and cloud fusion
The Share Model stored in shared platform is calculated to be exported as a result, place further is normalized to the output result
Reason, and Confidence evaluation is carried out to the output result with comentropy, and summation is weighted based on evaluation score and obtains label
Score, the fusion calculation module are special using the environment that the robot terminal and/or other described intelligent terminals upload
Reference breath and the label score of model normalizing unit output carry out fusion calculation and obtain the new Share Model.
Preferred embodiment according to the present invention, the cloud fusion calculation shared platform further include under data reception module, data
Module, model memory module are carried, the data reception module is used for whole with the robot terminal and/or other described intelligence
The privately owned model that the robot terminal and/or other described intelligent terminals upload received after the connection setup of end described
Cloud fusion calculation shared platform, the data download module are used for whole with the robot terminal and/or other described intelligence
End connection setup after by the Share Model in the cloud fusion calculation shared platform download to the robot terminal and/or
Other described intelligent terminals and/or other robot terminal, the model memory module are total for storing the cloud fusion calculation
Enjoy the Share Model of platform generation.
In order to solve the above technical problems, the present invention provides a kind of shared learning method of the robot based on cloud platform, including
Following steps:
Step 1: privately owned model generates terminal and collects environmental information, and generates local privately owned model；
Step 2: privately owned model generates terminal and the privately owned model locally generated and the environmental characteristic information of collection is uploaded cloud
Fusion calculation shared platform；
Step 3: cloud fusion calculation shared platform is by the privately owned model of upload, in conjunction with other robot terminal, privately owned model
It generates terminal and/or other privately owned models generates the environmental characteristic information and being total in cloud fusion calculation shared platform that terminal uploads
It enjoys model and carries out fusion calculation, generate new Share Model；
Step 4: other robot terminal, the privately owned model generate terminal and/or other privately owned models generate under terminal
Carry the Share Model in cloud fusion calculation shared platform.
Preferred embodiment according to the present invention, in step 1, the privately owned model generates the step that terminal generates local privately owned model
It is rapid as follows:
Step 1.1: the privately owned model generates the emulation structure that terminal collects environmental information or receives the input of other devices
The characteristic information made is as the environmental characteristic information；
Step 1.2: the privately owned model generates terminal and is based on the environmental characteristic use of information gazebo software building ring
Border model；
Step 1.3: running nitrification enhancement on the environmental model, obtain navigation strategy, the privately owned model packet
Include the navigation strategy and the environmental model.
Preferred embodiment according to the present invention, in step 2, it is described privately owned by what is locally generated that the privately owned model generates terminal
The step of model and the environmental characteristic information of collection upload the cloud fusion calculation shared platform is as follows:
Step 2.1: the privately owned model generates terminal and obtains the cloud fusion calculation shared platform address and send
Request；
Step 2.2: the privately owned model generates terminal and receives after the information of the cloud fusion calculation shared platform to institute
It states cloud fusion calculation shared platform and uploads the privately owned model and the environmental characteristic information.
Preferred embodiment according to the present invention, in step 3, the cloud fusion calculation shared platform generates new Share Model step
Suddenly include:
Step 3.1: the privately owned model is generated the environmental characteristic that terminal uploads and believed by the cloud fusion calculation shared platform
Breath be input in the Share Model of the privately owned model and the storage in the cloud fusion calculation shared platform exported as a result,
And the output result is normalized；
Step 3.2: the cloud fusion calculation shared platform does Confidence with result of the comentropy to the normalized
Evaluation；
Step 3.3: the cloud fusion calculation shared platform is weighted based on the evaluation score that the Confidence is evaluated to be asked
With obtain label score；
Step 3.4: the cloud fusion calculation shared platform generates the environmental characteristic that terminal uploads according to the privately owned model
Information and the label score carry out that new Share Model is calculated.
Preferred embodiment according to the present invention, in step 4, the other robot terminal, the privately owned model generate terminal
And/or transfer learning is also carried out after the Share Model in other privately owned models generation terminal downloads cloud fusion calculation shared platforms,
New privately owned model is generated, wherein the privately owned model, which generates terminal, carries out transfer learning step are as follows:
Step 4.1: the privately owned model generates terminal and downloads the Share Model from the cloud fusion calculation shared platform；
Step 4.2: the environmental characteristic data newly collected being input in the Share Model, the evaluation of all directions is exported
Value；
Step 4.3: institute's evaluation values being added in the environmental characteristic information, as new environmental model；
Step 4.4: intensified learning is carried out in new environmental model, generate new privately owned model for further upload to
The cloud fusion calculation shared platform.
In order to solve the above technical problems, the present invention provides a kind of method shared based on cloud fusion calculation, including following step
It is rapid:
Step 1: the robot terminal received, privately owned model are generated into the privately owned model and other robot that terminal generates
Terminal and/or other privately owned models generate the shared mould in the environmental characteristic information and cloud fusion calculation shared platform that terminal uploads
Type carries out fusion calculation, generates new Share Model, and the new Share Model is used to supply the robot terminal, other intelligence
Terminal and/or other robot terminal downloads and/or study.
Step 1 preferred embodiment according to the present invention generates new Share Model step are as follows:
Step 1.1: by the environmental characteristic information input received to the privately owned model and cloud fusion calculation shared platform
On Share Model in exported as a result, and the output result is normalized；
Step 1.2: doing Confidence evaluation with result of the comentropy to the normalized；
Step 1.3: the evaluation score based on Confidence evaluation is weighted summation and obtains label score；
Step 1.4: being carried out that the new shared mould is calculated according to the environmental characteristic information and the label score
Type.
Compared to the prior art, the present invention is based in the shared study of the robot of cloud platform and system, the cloud fusion is counted
The privately owned model that privately owned model generates terminal upload can be received by calculating shared platform, in conjunction with other robot terminal, privately owned model
It generates terminal and/or other privately owned models generates the environmental characteristic information and being total in cloud fusion calculation shared platform that terminal uploads
It enjoys model and carries out fusion calculation, generate new Share Model；So that other robot terminal, the privately owned model generate terminal
And/or other privately owned models generate the Share Model in terminal downloads cloud fusion calculation shared platform, use for local navigation, and
And can generate new privately owned model with further progress transfer learning and upload in the cloud fusion calculation shared platform, for into
One step downloading and shared, it is long to efficiently solve the training time in the training in robot navigation's decision model, can not carry out
The problem of experience merges.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is that the present invention is based on the system diagrams that learning system is shared by the robot of cloud platform.
Fig. 2 is that the present invention is based on the flow charts that learning system is shared by the robot of cloud platform.
Fig. 3 is that the present invention is based on the robots of cloud platform to share learning method flow chart.
Fig. 4 is that the present invention is based on the robots of cloud platform to share learning method transfer learning flow chart.
Fig. 5 is that the present invention is based on the robots of cloud platform to share learning method fusion calculation flow chart.
Fig. 6 is to realize that the electronics of learning method is shared by the robot based on cloud platform at least one example of the invention
The structural schematic diagram of equipment preferred embodiment.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
In order to make the foregoing objectives, features and advantages of the present invention clearer and more comprehensible, with reference to the accompanying drawing and specific real
Applying mode, the present invention is described in further detail.
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
The embodiment of a part of the invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people
The model that the present invention protects all should belong in member's every other embodiment obtained without making creative work
It encloses.
Description and claims of this specification and term " first " in above-mentioned attached drawing, " second " and " third " etc. are
For distinguishing different objects, not for description particular order.In addition, term " includes " and their any deformations, it is intended that
Non-exclusive include in covering.Such as the process, method, system, product or equipment for containing a series of steps or units do not have
It is defined in listed step or unit, but optionally further comprising the step of not listing or unit, or optionally further comprising
For the intrinsic other step or units of these process, methods, product or equipment.
As shown in Figure 1, Fig. 1 is to realize that the present invention is based on the system diagrams that learning system is shared by the robot of cloud platform.It is described
Learning system 100 is shared by robot based on cloud platform, including privately owned model generates terminal 110 and cloud fusion calculation shared platform
120.It includes characteristic collection module 111, environmental simulation module 112, reinforcing and transfer learning that the privately owned model, which generates terminal 110,
Module 113, the first data interpretation module 114, data uploading module 115 and first communication module 116 are learned wherein strengthening and migrating
Practising module 113 includes intensified learning unit 113a and migration computing unit 113b.The cloud fusion calculation shared platform 120 includes
Data download module 121, Model Fusion computing module 122, model memory module 123, the second data reception module 124 and second
Communication module 125, wherein Model Fusion computing module 122 includes model normalizing unit 122a and fusion calculation unit 122b.
Shown in referring to Figure 2 together, the present invention is based on the flow charts that learning system is shared by the robot of cloud platform by Fig. 2.Institute
It states privately owned model and generates characteristic collection module 111 described in terminal 110 for collecting environmental characteristic information, the environmental simulation mould
Block 112 is used to utilize the environmental characteristic information build environment model, includes strengthening in the reinforcing and transfer learning module 113
Unit 113a carries out intensified learning after inputting the environmental characteristic information on the environmental model, exports described privately owned
Model, the privately owned model include that the environmental model and the privately owned model generate terminal 110 for environmental model life
At navigation strategy.The privately owned model generates the first communication module 116 of terminal 110 by obtaining based on the cloud fusion
120 network address of shared platform is calculated, and is communicated with the foundation of the second communication module 124 of the cloud fusion calculation shared platform 120,
The data uploading module 115 is built for communicating in the first communication module 116 with the cloud fusion calculation shared platform 120
The privately owned model is uploaded into the cloud fusion calculation shared platform 120 after vertical.
Second data reception module 124 described in the cloud fusion calculation shared platform 120 be used for the privately owned mould
Type generate after 110 connection setup of terminal by the privately owned model generate the privately owned model that terminal 110 uploads receive it is described
Cloud fusion calculation shared platform 120.Model Fusion computing module 122 in the cloud fusion calculation shared platform 120 is used for will
The privately owned model uploaded generates terminal 110 and/or other privately owned moulds in conjunction with other robot terminal, the privately owned model
Type generate in the environmental characteristic information and the cloud fusion calculation shared platform 120 that terminal 110 uploads the Share Model that stores into
Row fusion calculation, generates new Share Model, and the new Share Model is used for for other robot terminal, the privately owned model
It generates terminal 110 and/or other privately owned models generates the downloading of terminal 120 and/or study.Wherein the cloud fusion calculation is shared flat
The model normalizing unit 122a in platform 120 is used to the environmental characteristic information input generating terminal to the privately owned model
The 110 privately owned models uploaded and the Share Model stored in the cloud fusion calculation shared platform 120 are exported
As a result further the output result is normalized, and Confidence is carried out to the output result with comentropy and is commented
Valence, and the evaluation score based on Confidence evaluation is weighted summation and obtains label score, the fusion calculation unit
122b is defeated using the environmental characteristic information of the privately owned model generation upload of terminal 110 and the model normalizing unit 122a
The label score out carries out fusion calculation and obtains the new Share Model.The privately owned model generates institute in terminal 110
It states the first data reception module 114 and the data download module 121 of the cloud fusion calculation shared platform 120 is logical in the two
After letter is established, the Share Model in the cloud fusion calculation shared platform 120 in the model memory module 123 is downloaded
Terminal 110 is generated to the privately owned model.The privately owned model generates the reinforcing and transfer learning module in terminal 110
The 113 migration computing unit 113b is migrated after by the Share Model of the environmental characteristic information input to downloading
It calculates, exports new environmental characteristic information, carry out intensified learning for the intensified learning unit 113a, to export new privately owned mould
Type, the privately owned model generate terminal 110 and the new privately owned model are also uploaded to the cloud fusion calculation shared platform 120
With further progress fusion calculation, new Share Model is generated, the new Share Model is used for for other robot terminal, institute
It states privately owned model and generates terminal 110 and/or other privately owned models generation terminal 110 downloadings and/or study.By repeating this step
Suddenly, the model on cloud becomes stronger and stronger.
Specifically, the privately owned model generates terminal 110 using intelligent robot terminal, computer terminal and/or other intelligence
It can terminal device.Wherein the intelligent robot terminal can use the mobile robot of products storage circulation system, and the environment is special
Reference breath can be the environmental characteristic information collected in warehouse logistics environment, but be not limited with above-mentioned, the privately owned model
Generate terminal 110 or other navigating robots applied to a variety of environment.The feature collection module 111 is using camera shooting
The spy that environmental characteristic information described in head and/or laser radar random acquisition or the emulation for directly receiving the input of other devices construct
Reference breath is used as the environmental characteristic information.The environmental simulation module constructs simulated environment using gazebo simulation software.
As shown in figure 3, Fig. 3 is that the present invention is based on the streams that the learning method that learning system uses is shared by the robot of cloud platform
Cheng Tu.It is appreciated that the sequence of step can change in the flow chart according to different requirements, certain steps be can be omitted.
Step S1: privately owned model generates terminal and collects environmental information, and generates local privately owned model；
Step S2: privately owned model generates terminal and the privately owned model locally generated and the environmental characteristic information of collection is uploaded cloud
Fusion calculation shared platform；
Step S3: cloud fusion calculation shared platform is by the privately owned model of upload, in conjunction with other robot terminal, privately owned model
It generates terminal and/or other privately owned models generates the environmental characteristic information and being total in cloud fusion calculation shared platform that terminal uploads
It enjoys model and carries out fusion calculation, generate new Share Model；
Step S4: other robot terminal, the privately owned model generate terminal and/or other privately owned models generate under terminal
The Share Model in cloud fusion calculation shared platform is carried, transfer learning is carried out, generates new privately owned model.
Specifically, transfer learning in step S4 is further described as follows.
As shown in figure 4, Fig. 4 is that the present invention is based on the robots of cloud platform to share learning method transfer learning flow chart:
Step S41: privately owned model generates terminal and collects environmental information or receive the emulation construction of other devices input
Characteristic information is run strong as the environmental characteristic use of information gazebo software building environmental model on the environmental model
Change learning algorithm, obtains navigation strategy, the privately owned model includes the navigation strategy and the environmental model.The privately owned mould
Type generates terminal and shares the privately owned model locally generated and the environmental characteristic information of the collection upload cloud fusion calculation
Platform generates 1st generation Share Model.
Step S42: the privately owned model generates terminal and downloads the Share Model from the cloud fusion calculation shared platform,
The environmental characteristic data newly collected are input in the Share Model, the output of Share Model is output to as supplementary features
In Q network, or all parameters are transmitted in Q network, export the evaluation of estimate of all directions, institute's evaluation values is added to described
In environmental characteristic information, as new environmental model；
Step S43: carrying out intensified learning in new environmental model, the feature vector of input layer by original feature vector and
The vector composition of Share Model output, it is shared flat to the cloud fusion calculation for further uploading to export new privately owned model
Platform generates 2nd generation Share Model.Further, about the 3rd generation Share Model, the 4th generation Share Model and the n-th generation Share Model
Generating principle and 2nd generation Share Model generating principle it is essentially identical, be not described in more detail here.
Further, fusion calculation in step S3 is further described as follows.
As shown in figure 5, Fig. 5 is that the present invention is based on the robots of cloud platform to share the fusion calculation method that learning platform uses
Flow chart.Step S3 may include step S31, S32 and S33.
Step S31: the privately owned model is generated the environmental characteristic that terminal uploads and believed by the cloud fusion calculation shared platform
Breath, which is input in the Share Model of the privately owned model and the storage in the cloud fusion calculation shared platform, obtains output result.
Step S32: being normalized the output result, the cloud fusion calculation shared platform comentropy pair
The result of the normalized does Confidence evaluation, and the cloud fusion calculation shared platform is commented based on what the Confidence was evaluated
Valence score is weighted summation and obtains label score.
Step S33: the cloud fusion calculation shared platform generates the environmental characteristic that terminal uploads according to the privately owned model
Information and the label score carry out that new Share Model is calculated.
Compared to the prior art, the present invention is based in the shared study of the robot of cloud platform and system, the cloud fusion is counted
The privately owned model that privately owned model generates terminal upload can be received by calculating shared platform, in conjunction with other robot terminal, privately owned model
It generates terminal and/or other privately owned models generates the environmental characteristic information and being total in cloud fusion calculation shared platform that terminal uploads
It enjoys model and carries out fusion calculation, generate new Share Model；So that other robot terminal, the privately owned model generate terminal
And/or other privately owned models generate the Share Model in terminal downloads cloud fusion calculation shared platform, use for local navigation, and
And can generate new privately owned model with further progress transfer learning and upload in the cloud fusion calculation shared platform, for into
One step downloading and shared, it is long to efficiently solve the training time in the training in robot navigation's decision model, can not carry out
The problem of experience merges.
As shown in fig. 6, to execute the structural schematic diagram of the computer installation 5 of the method in above-described embodiment.The calculating
Machine device 5 includes, but are not limited to: at least one processor 51, at least one processor 52, at least one communication device 53 and
At least one communication bus.Wherein, the communication bus is for realizing the connection communication between these components.
The computer installation 5 be it is a kind of can according to the instruction for being previously set or store, it is automatic carry out numerical value calculate with/
Or the equipment of information processing, hardware include but is not limited to microprocessor, specific integrated circuit (Application Specific
Integrated Circuit, ASIC), programmable gate array (Field-Programmable Gate Array, FPGA), number
Word processing device (Digital Signal Processor, DSP), embedded device etc..The computer installation 5 may also include net
Network equipment and/or user equipment.Wherein, the network equipment includes but is not limited to single network server, multiple network services
The server group of device composition or being made of a large amount of hosts or network server based on cloud computing (Cloud Computing)
Cloud, wherein cloud computing is one kind of distributed computing, a super virtual meter consisting of a loosely coupled set of computers
Calculation machine.
The computer installation 5 may be, but not limited to, any one and can be set with user by keyboard, touch tablet or acoustic control
The modes such as standby carry out the electronic product of human-computer interaction, for example, tablet computer, smart phone, personal digital assistant (Personal
Digital Assistant, PDA), intellectual wearable device, picture pick-up device, the terminals such as monitoring device.
Network locating for the computer installation 5 includes, but are not limited to internet, wide area network, Metropolitan Area Network (MAN), local area network, void
Quasi- dedicated network (Virtual Private Network, VPN) etc..
Wherein, the communication device 53 can be wired sending port, or and wireless device is traditional thread binding for example including day
It sets, for carrying out data communication with other equipment.
The memory 51 is for storing program code.The memory 51, which can be, does not have physical form in integrated circuit
The circuit with store function, such as RAM (Random-Access Memory, random access memory), FIFO (First In
First Out) etc..Alternatively, the memory is also possible to the memory with physical form, such as memory bar, TF card
(Trans-flash Card), smart media card (smart media card), safe digital card (secure digital
Card), storage facilities such as flash memory cards (flash card) etc..
The processor 52 may include one or more microprocessor, digital processing unit.The processor can call
The program code stored in the memory is to execute relevant function；For example, modules described in Fig. 3 are stored in and deposit
The program code of reservoir, and as performed by the processor, to realize that learning method is shared by a kind of cloud robot.The processor
Also known as central processing unit (CPU, Central Processing Unit), is one piece of ultra-large integrated circuit, is operation core
The heart (Core) and control core (Control Unit).
The embodiment of the present invention also provides a kind of computer readable storage medium, is stored thereon with computer instruction, the finger
It enables when being included that one or more processors execute, the robot learning based on cloud platform is made to execute such as embodiment of the method above
Learning method is shared by the cloud robot.
The characteristic means of present invention mentioned above can be realized by integrated circuit, and control above-mentioned of realization
The function of learning method is shared by cloud robot described in embodiment of anticipating.
Function achieved by the shared learning method of the cloud robot described in any embodiment can be transferred through of the invention
Integrated circuit is installed in the electronic equipment, and playing the electronic equipment, cloud robot described in any embodiment is shared to be learned
Function achieved by learning method, this will not be detailed here.
It should be noted that for the various method embodiments described above, for simple description, therefore, it is stated as a series of
Combination of actions, but those skilled in the art should understand that, the present invention is not limited by the sequence of acts described because
According to the present invention, some steps may be performed in other sequences or simultaneously.Secondly, those skilled in the art should also know
It knows, the embodiments described in the specification are all preferred embodiments, and related actions and modules is not necessarily of the invention
It is necessary.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, there is no the portion being described in detail in some embodiment
Point, reference can be made to the related descriptions of other embodiments.
In several embodiments provided herein, it should be understood that disclosed device, it can be by another way
It realizes.For example, the apparatus embodiments described above are merely exemplary, such as the division of the unit, it is only a kind of
Logical function partition, there may be another division manner in actual implementation, such as multiple units or components can combine or can
To be integrated into another system, or some features can be ignored or not executed.Another point, shown or discussed is mutual
Coupling, direct-coupling or communication connection can be through some interfaces, the indirect coupling or communication connection of device or unit,
It can be electrical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
In addition, each functional unit in various embodiments of the present invention can integrate in one processing unit, it can also
To be that each unit physically exists alone, can also be integrated in one unit with two or more units.It is above-mentioned integrated
Unit both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially
The all or part of the part that contributes to existing technology or the technical solution can be in the form of software products in other words
It embodies, which is stored in a storage medium, including some instructions are used so that a computer
Equipment (can for personal computer, server or network equipment etc.) execute each embodiment the method for the present invention whole or
Part steps.And storage medium above-mentioned includes: that USB flash disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
Reservoir (RAM, Random Access Memory), mobile hard disk, magnetic or disk etc. be various to can store program code
Medium.The above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although referring to aforementioned
Invention is explained in detail for embodiment, those skilled in the art should understand that: it still can be to aforementioned
Technical solution documented by each embodiment is modified or equivalent replacement of some of the technical features；And these are repaired
Change or replaces, the range for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution.
Claims (20)
1. learning system is shared by a kind of robot based on cloud platform comprising privately owned model generates terminal and cloud fusion calculation is total
Enjoy platform, it is characterised in that: the privately owned model generates terminal and is used for the privately owned model locally generated and is used to generate described
Privately owned model and the environmental characteristic information collected uploads the cloud fusion calculation shared platform, the cloud fusion calculation shared platform
Including Model Fusion computing module, the Model Fusion computing module is used for the privately owned model that will be uploaded, in conjunction with other machines
Device people terminal, the privately owned model generate terminal and/or other privately owned models generate the environmental characteristic information and institute that terminal uploads
It states the Share Model stored in cloud fusion calculation shared platform and carries out fusion calculation, generate new Share Model, described new is total to
Model is enjoyed for generating terminal and/or other privately owned models generation terminal downloads for other robot terminal, the privately owned model
And/or study.
2. learning system is shared by the robot according to claim 1 based on cloud platform, it is characterised in that: the privately owned mould
It includes characteristic collection module, environmental simulation module, reinforcing and transfer learning module that type, which generates terminal, and the characteristic collection module is used
In collecting environmental characteristic information, the environmental simulation module is used to utilize the environmental characteristic information build environment model, described
Strengthen and transfer learning module includes intensified learning unit, the intensified learning unit on the environmental model for inputting institute
State and carry out intensified learning after environmental characteristic information, export the privately owned model, the privately owned model include the environmental model and
The privately owned model generates terminal and is directed to the navigation strategy that the environmental model generates.
3. learning system is shared by the robot according to claim 2 based on cloud platform, it is characterised in that: it is described reinforcing and
Transfer learning module further includes migration computing unit, and the migration computing unit by the environmental characteristic information input for arriving
After the Share Model of downloading, migration calculating is carried out, exports new environmental characteristic information, is strengthened for the intensified learning unit
Study, to export the new privately owned model, the privately owned model generates terminal and the new privately owned model is also uploaded to institute
Cloud fusion calculation shared platform is stated with further progress fusion calculation.
4. learning system is shared by the robot according to claim 1 or 2 based on cloud platform, it is characterised in that: the private
There is model to generate terminal using intelligent robot terminal, computer terminal or other intelligent terminals.
5. learning system is shared by the robot according to claim 2 based on cloud platform, it is characterised in that: the feature is received
Collection module acquires the environmental characteristic information using camera and/or laser radar or directly receives the imitative of other devices input
The characteristic information really constructed is as the environmental characteristic information.
6. learning system is shared by the robot according to claim 2 based on cloud platform, it is characterised in that: the environment is imitative
True module constructs simulated environment using gazebo simulation software.
7. learning system is shared by the robot according to claim 1 based on cloud platform, it is characterised in that: the model melts
Closing computing module includes model normalizing unit, fusion calculation unit, and the model normalizing unit is for believing the environmental characteristic
Breath is input to the privately owned model that the privately owned model generation terminal uploads and stores in the cloud fusion calculation shared platform
The Share Model obtain output result further the output result be normalized, and with comentropy to described
It exports result and carries out Confidence evaluation, and the evaluation score based on Confidence evaluation is weighted summation and obtains label point
Number, the fusion calculation unit generates the environmental characteristic information of terminal upload using the privately owned model and the model is returned
The label score of the output of Unit one carries out fusion calculation and obtains the new Share Model.
8. learning system is shared by the robot according to claim 1 based on cloud platform, it is characterised in that: the privately owned mould
It further includes first communication module, the first data reception module, data uploading module that type, which generates terminal, and the cloud fusion calculation is shared
Platform further includes second communication module, and the first communication module is for obtaining the cloud fusion calculation shared platform network
Location, and communicated with the foundation of the second communication module of the cloud fusion calculation shared platform, the data uploading module is used in institute
First communication module is stated to merge the privately owned model upload cloud with after the cloud fusion calculation shared platform connection setup
Calculate shared platform, first data reception module be used for after the cloud fusion calculation shared platform connection setup by institute
It states the Share Model in cloud fusion calculation shared platform and downloads to the privately owned model generation terminal.
9. learning system is shared by the robot according to claim 1 based on cloud platform, it is characterised in that: the cloud fusion
Calculating shared platform further includes the second data reception module, data download module, model memory module, second data receiver
Module is used to generate the private for uploading the privately owned model generation terminal after terminal connection setup with the privately owned model
There is model to receive the cloud fusion calculation shared platform, the data download module is used to generate eventually with the privately owned model
The Share Model in the cloud fusion calculation shared platform the privately owned model is downloaded to after the connection setup of end to generate eventually
End, the model memory module are used to store the Share Model in the cloud fusion calculation shared platform.
10. one kind is based on cloud fusion calculation shared platform, it is characterised in that: the cloud fusion calculation shared platform includes that model melts
Computing module is closed, the Model Fusion computing module is used for the privately owned mould for uploading robot terminal and/or other intelligent terminals
Share Model in type, environmental characteristic information and the cloud fusion calculation shared platform carries out fusion calculation, generates new share
Model, the new Share Model are used for for the robot terminal and/or other intelligent terminals and/or other robot terminal
Downloading and/or study.
11. according to claim 10 be based on cloud fusion calculation shared platform, it is characterised in that: the Model Fusion calculates
Module includes model normalizing unit, fusion calculation unit, and the model normalizing unit is used for the robot terminal and/or institute
The environmental characteristic information input of other intelligent terminals upload is stated to the robot terminal and/or other described intelligent terminals
On the privately owned model uploaded and the cloud fusion calculation shared platform Share Model that stores exported as a result, into
The output result is normalized in one step, and carries out Confidence evaluation, and base to the output result with comentropy
Summation is weighted in evaluation score and obtains label score, and the fusion calculation module utilizes the robot terminal and/or institute
The label score of the environmental characteristic information and model normalizing unit output of stating the upload of other intelligent terminals carries out
Fusion calculation obtains the new Share Model.
12. according to claim 10 be based on cloud fusion calculation shared platform, it is characterised in that: the cloud fusion calculation is total
Enjoying platform further includes data reception module, data download module, model memory module, the data reception module be used for institute
It states the robot terminal and/or other described intelligence after robot terminal and/or other described intelligent terminal connection setups
The privately owned model that terminal uploads receives the cloud fusion calculation shared platform, the data download module be used for institute
Stating will be described total in the cloud fusion calculation shared platform after robot terminal and/or other described intelligent terminal connection setups
Enjoy that model downloads to the robot terminal and/or other described intelligent terminals and/or other robot terminal, the model are deposited
Storage module is used to store the Share Model that the cloud fusion calculation shared platform generates.
13. learning method is shared by a kind of robot based on cloud platform, which comprises the steps of:
Step 1: privately owned model generates terminal and collects environmental information, and generates local privately owned model；
Step 2: privately owned model generates terminal and the privately owned model locally generated and the environmental characteristic information of collection is uploaded cloud fusion
Calculate shared platform；
Step 3: cloud fusion calculation shared platform generates the privately owned model of upload in conjunction with other robot terminal, privately owned model
Terminal and/or other privately owned models generate the shared mould in the environmental characteristic information and cloud fusion calculation shared platform that terminal uploads
Type carries out fusion calculation, generates new Share Model；
Step 4: other robot terminal, the privately owned model generate terminal and/or other privately owned models generate terminal downloads cloud
Share Model in fusion calculation shared platform.
14. learning method is shared by the robot according to claim 13 based on cloud platform, it is characterised in that: the step
In 1, it is as follows that the privately owned model generates the step of terminal generates local privately owned model:
Step 1.1: the privately owned model generates terminal and collects environmental information or receive the emulation construction of other devices input
Characteristic information is as the environmental characteristic information；
Step 1.2: the privately owned model generates terminal and is based on the environmental characteristic use of information gazebo software building environment mould
Type；
Step 1.3: running nitrification enhancement on the environmental model, obtain navigation strategy, the privately owned model includes institute
State navigation strategy and the environmental model.
15. learning method is shared by the robot according to claim 13 based on cloud platform, it is characterised in that: in step 2,
The privately owned model generates terminal and melts the privately owned model locally generated and the environmental characteristic information of the collection upload cloud
Total the step of calculating shared platform, is as follows:
Step 2.1: the privately owned model generates terminal and obtains the cloud fusion calculation shared platform address and send request；
Step 2.2: the privately owned model generates terminal and receives Xiang Suoshu cloud after the information of the cloud fusion calculation shared platform
Fusion calculation shared platform uploads the privately owned model and the environmental characteristic information.
16. learning method is shared by the robot according to claim 13 based on cloud platform, it is characterised in that: the step
In 3, the cloud fusion calculation shared platform generates new Share Model step and includes:
Step 3.1: the environmental characteristic information that the privately owned model is generated terminal upload by the cloud fusion calculation shared platform is defeated
Enter into the Share Model of the storage on the privately owned model and the cloud fusion calculation shared platform and is exported as a result, and right
The output result is normalized；
Step 3.2: the cloud fusion calculation shared platform does Confidence evaluation with result of the comentropy to the normalized；
Step 3.3: the cloud fusion calculation shared platform is weighted based on the evaluation score that the Confidence is evaluated sums
To label score；
Step 3.4: the cloud fusion calculation shared platform generates the environmental characteristic information that terminal uploads according to the privately owned model
With the label score carry out that new Share Model is calculated.
17. learning method is shared by the robot according to claim 13 based on cloud platform, it is characterised in that: the step
In 4, the other robot terminal, the privately owned model generate terminal and/or other privately owned models generate terminal downloads cloud and melt
Transfer learning is also carried out after total Share Model calculated in shared platform, new privately owned model is generated, wherein the privately owned model
It generates terminal and carries out transfer learning step are as follows:
Step 4.1: the privately owned model generates terminal and downloads the Share Model from the cloud fusion calculation shared platform；
Step 4.2: the environmental characteristic data newly collected being input in the Share Model, the evaluation of estimate of all directions is exported；
Step 4.3: institute's evaluation values being added in the environmental characteristic information, as new environmental model；
Step 4.4: carrying out intensified learning in new environmental model, generate new privately owned model for further uploading to described
Cloud fusion calculation shared platform.
18. a kind of method shared based on cloud fusion calculation, which comprises the following steps:
Step 1: the robot terminal received, privately owned model are generated into the privately owned model and other robot terminal that terminal generates
And/or other privately owned models generate the Share Model in the environmental characteristic information and cloud fusion calculation shared platform that terminals upload into
Row fusion calculation, generates new Share Model, and the new Share Model is used to supply the robot terminal, other intelligent terminals
And/or other robot terminal downloads and/or study.
19. the method according to claim 18 shared based on cloud fusion calculation, it is characterised in that: step 1, generate new
Share Model step are as follows:
Step 1.1: will be in the environmental characteristic information input that received to the privately owned model and cloud fusion calculation shared platform
It is exported in Share Model as a result, and the output result is normalized；
Step 1.2: doing Confidence evaluation with result of the comentropy to the normalized；
Step 1.3: the evaluation score based on Confidence evaluation is weighted summation and obtains label score；
Step 1.4: according to the environmental characteristic information and the label score carrying out that the new Share Model is calculated.
20. a kind of computer readable storage medium, which is characterized in that the computer-readable recording medium storage has at least one
Instruction, at least one described instruction are realized when being executed by processor such as institute in any one of claim 13-19 claim
The method stated.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910248301.9A CN109993308A (en) | 2019-03-29 | 2019-03-29 | Learning system and method, shared platform and method, medium are shared based on cloud platform |
| PCT/CN2019/130567 WO2020199690A1 (en) | 2019-03-29 | 2019-12-31 | Cloud platform-based sharing learning system and method, sharing platform and method, and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910248301.9A CN109993308A (en) | 2019-03-29 | 2019-03-29 | Learning system and method, shared platform and method, medium are shared based on cloud platform |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109993308A true CN109993308A (en) | 2019-07-09 |
Family
ID=67131830
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910248301.9A Pending CN109993308A (en) | 2019-03-29 | 2019-03-29 | Learning system and method, shared platform and method, medium are shared based on cloud platform |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN109993308A (en) |
| WO (1) | WO2020199690A1 (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110766169A (en) * | 2019-10-31 | 2020-02-07 | 深圳前海微众银行股份有限公司 | Transfer training optimization method and device for reinforcement learning, terminal and storage medium |
| CN110796267A (en) * | 2019-11-12 | 2020-02-14 | 支付宝(杭州)信息技术有限公司 | Machine learning method and machine learning device for data sharing |
| CN110827167A (en) * | 2019-09-29 | 2020-02-21 | 武汉开目信息技术股份有限公司 | Product design manufacturability knowledge sharing method and device for collaborative manufacturing |
| CN111027713A (en) * | 2019-12-10 | 2020-04-17 | 支付宝(杭州)信息技术有限公司 | Shared machine learning system and method |
| WO2020199690A1 (en) * | 2019-03-29 | 2020-10-08 | 深圳先进技术研究院 | Cloud platform-based sharing learning system and method, sharing platform and method, and medium |
| CN112100145A (en) * | 2020-09-02 | 2020-12-18 | 南京三眼精灵信息技术有限公司 | Digital model sharing learning system and method |
| WO2021120951A1 (en) * | 2019-12-20 | 2021-06-24 | 深圳前海微众银行股份有限公司 | Knowledge transfer method, apparatus and device based on federated learning, and medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103926838A (en) * | 2014-04-22 | 2014-07-16 | 山东大学 | Autonomous metal development cloud robot system based on cloud computing |
| US20160067864A1 (en) * | 2014-09-05 | 2016-03-10 | Accenture Global Services Limited | Self-adaptive device intelligence as a service enterprise infrastructure for sensor-rich environments |
| CN106027300A (en) * | 2016-05-23 | 2016-10-12 | 深圳市飞仙智能科技有限公司 | System and method for parameter optimization of intelligent robot applying neural network |
| US20160311115A1 (en) * | 2015-04-27 | 2016-10-27 | David M. Hill | Enhanced configuration and control of robots |
| CN108398660A (en) * | 2018-01-08 | 2018-08-14 | 国网江苏省电力有限公司 | A kind of terminal device localization method and system based on Wi-Fi cloud platform systems |
| CN108896739A (en) * | 2018-07-16 | 2018-11-27 | 河南聚合科技有限公司 | It is a kind of can assaying ingredient solar-energy machine people's cloud platform |
| CN109116854A (en) * | 2018-09-16 | 2019-01-01 | 南京大学 | A kind of robot cooperated control method of multiple groups based on intensified learning and control system |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107053184A (en) * | 2017-06-22 | 2017-08-18 | 北京信息科技大学 | Multi-Agent Cooperation processing system and method based on private clound |
| CN108921298B (en) * | 2018-06-12 | 2022-04-19 | 中国科学技术大学 | Multi-agent communication and decision-making method for reinforcement learning |
| CN109086550B (en) * | 2018-08-27 | 2019-05-28 | 山东师范大学 | The evacuation emulation method and system of Q study are shared based on multi-Agent |
| CN109993308A (en) * | 2019-03-29 | 2019-07-09 | 深圳先进技术研究院 | Learning system and method, shared platform and method, medium are shared based on cloud platform |
-
2019
- 2019-03-29 CN CN201910248301.9A patent/CN109993308A/en active Pending
- 2019-12-31 WO PCT/CN2019/130567 patent/WO2020199690A1/en active Application Filing
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103926838A (en) * | 2014-04-22 | 2014-07-16 | 山东大学 | Autonomous metal development cloud robot system based on cloud computing |
| US20160067864A1 (en) * | 2014-09-05 | 2016-03-10 | Accenture Global Services Limited | Self-adaptive device intelligence as a service enterprise infrastructure for sensor-rich environments |
| US20160311115A1 (en) * | 2015-04-27 | 2016-10-27 | David M. Hill | Enhanced configuration and control of robots |
| CN106027300A (en) * | 2016-05-23 | 2016-10-12 | 深圳市飞仙智能科技有限公司 | System and method for parameter optimization of intelligent robot applying neural network |
| CN108398660A (en) * | 2018-01-08 | 2018-08-14 | 国网江苏省电力有限公司 | A kind of terminal device localization method and system based on Wi-Fi cloud platform systems |
| CN108896739A (en) * | 2018-07-16 | 2018-11-27 | 河南聚合科技有限公司 | It is a kind of can assaying ingredient solar-energy machine people's cloud platform |
| CN109116854A (en) * | 2018-09-16 | 2019-01-01 | 南京大学 | A kind of robot cooperated control method of multiple groups based on intensified learning and control system |
Non-Patent Citations (3)
| Title |
|---|
| HYUN KIM等: ""Client/Server Framework for Providing Context-Aware Services to Network Based Robots"", RO-MAN 2007 - THE 16TH IEEE INTERNATIONAL SYMPOSIUM ON ROBOT AND HUMAN INTERACTIVE COMMUNICATION * |
| TLIJANI HAYET等: ""A navigation model for a multi-robot system Based on Client/Server model"", 2016 INTERNATIONAL CONFERENCE ON CONTROL, DECISION AND INFORMATION TECHNOLOGIES (CODIT) * |
| 洪臣等: "\"云机器人架构和特征概述\"", 《机器人技术与应用》, no. 6 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020199690A1 (en) * | 2019-03-29 | 2020-10-08 | 深圳先进技术研究院 | Cloud platform-based sharing learning system and method, sharing platform and method, and medium |
| CN110827167A (en) * | 2019-09-29 | 2020-02-21 | 武汉开目信息技术股份有限公司 | Product design manufacturability knowledge sharing method and device for collaborative manufacturing |
| CN110766169A (en) * | 2019-10-31 | 2020-02-07 | 深圳前海微众银行股份有限公司 | Transfer training optimization method and device for reinforcement learning, terminal and storage medium |
| CN110796267A (en) * | 2019-11-12 | 2020-02-14 | 支付宝(杭州)信息技术有限公司 | Machine learning method and machine learning device for data sharing |
| CN111027713A (en) * | 2019-12-10 | 2020-04-17 | 支付宝(杭州)信息技术有限公司 | Shared machine learning system and method |
| WO2021120951A1 (en) * | 2019-12-20 | 2021-06-24 | 深圳前海微众银行股份有限公司 | Knowledge transfer method, apparatus and device based on federated learning, and medium |
| CN112100145A (en) * | 2020-09-02 | 2020-12-18 | 南京三眼精灵信息技术有限公司 | Digital model sharing learning system and method |
| CN112100145B (en) * | 2020-09-02 | 2023-07-04 | 南京三眼精灵信息技术有限公司 | Digital model sharing learning system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2020199690A1 (en) | 2020-10-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109993308A (en) | Learning system and method, shared platform and method, medium are shared based on cloud platform | |
| JP7159458B2 (en) | Method, apparatus, device and computer program for scheduling virtual objects in a virtual environment | |
| EP3992857A1 (en) | Method and device for generating neural network model, and computer-readable storage medium | |
| CN108463273A (en) | Mobile history based on player carries out the games system etc. of the path finding of non-gaming person role | |
| CN109782600A (en) | A method of autonomous mobile robot navigation system is established by virtual environment | |
| CN110523081A (en) | The method and device for planning in navigation pathfinding path | |
| CN110321666A (en) | Multi-robots Path Planning Method based on priori knowledge Yu DQN algorithm | |
| CN110339569A (en) | Control the method and device of virtual role in scene of game | |
| CN109464803A (en) | Virtual objects controlled, model training method, device, storage medium and equipment | |
| Jain et al. | DCOPs meet the real world: Exploring unknown reward matrices with applications to mobile sensor networks | |
| CN110516389B (en) | Behavior control strategy learning method, device, equipment and storage medium | |
| JP2023502860A (en) | Information processing method, device, computer program and electronic device | |
| CN110368688A (en) | Display methods, device, storage medium and the electronic device of animation | |
| CN110251942A (en) | Control the method and device of virtual role in scene of game | |
| CN114139637B (en) | Multi-agent information fusion method and device, electronic equipment and readable storage medium | |
| CN110132282A (en) | Unmanned plane paths planning method and device | |
| CN115300910B (en) | Confusion-removing game strategy model generation method based on multi-agent reinforcement learning | |
| CN113962390B (en) | Method for constructing diversified search strategy model based on deep reinforcement learning network | |
| WO2023024762A1 (en) | Artificial intelligence object control method and apparatus, device, and storage medium | |
| CN111897327A (en) | Multi-mobile-robot control/assignment model acquisition method and device and electronic equipment | |
| CN113509726B (en) | Interaction model training method, device, computer equipment and storage medium | |
| CN109529358A (en) | Feature integration method and apparatus and electronic device | |
| CN115797517B (en) | Data processing method, device, equipment and medium of virtual model | |
| CN109731338A (en) | Artificial intelligence training method and device, storage medium and electronic device in game | |
| CN113378656A (en) | Action identification method and device based on self-adaptive graph convolution neural network |
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 |