CN106773659A - A kind of robot learning by imitation method based on Gaussian process - Google Patents
A kind of robot learning by imitation method based on Gaussian process Download PDFInfo
- Publication number
- CN106773659A CN106773659A CN201510805218.9A CN201510805218A CN106773659A CN 106773659 A CN106773659 A CN 106773659A CN 201510805218 A CN201510805218 A CN 201510805218A CN 106773659 A CN106773659 A CN 106773659A
- Authority
- CN
- China
- Prior art keywords
- robot
- gaussian process
- imitation
- learning
- sample point
- 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
Abstract
The invention discloses a kind of robot learning by imitation method based on Gaussian process, comprise the following steps：The first step：Light source position arbitrarily sets；Second step：Teaching robot carries out action modeling, completes light action；3rd step：Sample point set is trained with the method for Gaussian process, sets up and solve its Gaussian process model；4th step：Imitate robot and this mapping relations is applied to itself control strategy, imitate the learning by imitation of teaching robot's behavior；5th step：Appropriation is analyzed.Robot learning by imitation method based on Gaussian process of the invention, Gaussian process is applied to the research of robot learning by imitation control strategy algorithm, by the teaching behavior sample point for gathering teaching robot, sample point training is carried out using Gaussian process algorithm, learn the mapping relations between the perception and behavior of teaching robot, and this mapping relations is applied to imitate robot learning by imitation control strategy and performed.
Description
Technical field
The present invention relates to a kind of robot learning by imitation method, and in particular to a kind of robot learning by imitation method based on Gaussian process, belong to intelligent artifact technical field.
Background technology
The mankind and other biological can be by observation and the behaviors of imitation companion, effectively obtain the motion mode of companion's behavior, and understand the connotation of these behaviors, assign the ability of robot learning by imitation as the mankind, the problem of robot motion's technical ability acquisition can be efficiently solved, the learning efficiency and adaptive ability of robot are improved, is an important research direction of robot bionic research；The general study form of learning by imitation is replicated for action, action replicates the movement locus for being mainly based upon demonstrator, by the regression problem for solving to be acted from demonstrator track to execution, to determine control strategy, imitator performs control strategy and carries out behavior reproduction, realize learning by imitation, control strategy is sought using inverse nitrification enhancement mostly in learning by imitation algorithm, the method of inverse intensified learning is mainly based upon imitator carries out the cost function of learning by imitation to teaching behavior, seek to make the minimum control strategy of cost function but, the method of inverse intensified learning is higher to cost function requirement, it is not suitable for the learning by imitation task that cost function is difficult to obtain.
The content of the invention
（One）The technical problem to be solved
To solve the above problems, the present invention proposes a kind of robot learning by imitation method based on Gaussian process, Gaussian process is applied to the research of robot learning by imitation control strategy algorithm, by the teaching behavior sample point for gathering teaching robot, sample point training is carried out using Gaussian process algorithm, learn the mapping relations between the perception and behavior of teaching robot, and this mapping relations is applied to imitate robot learning by imitation control strategy and performed.
（Two）Technical scheme
Robot learning by imitation method based on Gaussian process of the invention, comprises the following steps：
The first step：Teaching robot uses the non-crossing connected mode of Braitenberg cars, optical sensor output valve and the corresponding inversely proportional relation of motor output valve, light source position arbitrarily sets, robot is imitated equally using the non-crossing connected mode of Braitenberg cars, relation between optical sensor output valve and corresponding motor output valve is unknown, it is necessary to the strategy that learns by imitation is given；
Second step：Teaching robot carries out action modeling, completes light action, while randomly selecting sample point, constitutes sample point set, and each sample point includes two parameters；
3rd step：Sample point set is trained with the method for Gaussian process, sets up and solve its Gaussian process model, obtain the mapping relations between teaching robot's sensor and motor；
4th step：Imitate robot and this mapping relations is applied to itself control strategy, imitate the learning by imitation of teaching robot's behavior；
5th step：Appropriation is analyzed.
Further, two parameters in the second step are optical sensor output valve and corresponding motor output valve.
（Three）Beneficial effect
Compared with prior art, robot learning by imitation method based on Gaussian process of the invention, Gaussian process is applied to the research of robot learning by imitation control strategy algorithm, by the teaching behavior sample point for gathering teaching robot, sample point training is carried out using Gaussian process algorithm, learn the mapping relations between the perception and behavior of teaching robot, and this mapping relations is applied to imitate robot learning by imitation control strategy and performed.
Specific embodiment
A kind of robot learning by imitation method based on Gaussian process, comprises the following steps：
The first step：Teaching robot uses the non-crossing connected mode of Braitenberg cars, optical sensor output valve and the corresponding inversely proportional relation of motor output valve, light source position arbitrarily sets, robot is imitated equally using the non-crossing connected mode of Braitenberg cars, relation between optical sensor output valve and corresponding motor output valve is unknown, it is necessary to the strategy that learns by imitation is given；
Second step：Teaching robot carries out action modeling, completes light action, while randomly selecting sample point, constitutes sample point set, and each sample point includes two parameters；
3rd step：Sample point set is trained with the method for Gaussian process, sets up and solve its Gaussian process model, obtain the mapping relations between teaching robot's sensor and motor；
4th step：Imitate robot and this mapping relations is applied to itself control strategy, imitate the learning by imitation of teaching robot's behavior；
5th step：Appropriation is analyzed.
Wherein, two parameters in the second step are optical sensor output valve and corresponding motor output valve.
Embodiment described above is only that the preferred embodiment of the present invention is described, and not the spirit and scope of the present invention are defined.On the premise of design concept of the present invention is not departed from; the all variations and modifications that this area ordinary person makes to technical scheme; protection scope of the present invention all should be dropped into, claimed technology contents of the invention have all been recorded in detail in the claims.
Claims (2)
1. a kind of robot learning by imitation method based on Gaussian process, it is characterised in that：Comprise the following steps：
The first step：Teaching robot uses the non-crossing connected mode of Braitenberg cars, optical sensor output valve arbitrarily to be set with the corresponding inversely proportional relation of motor output valve, light source position；
Second step：Teaching robot carries out action modeling, completes light action, while randomly selecting sample point, constitutes sample point set, and each sample point includes two parameters；
3rd step：Sample point set is trained with the method for Gaussian process, sets up and solve its Gaussian process model, obtain the mapping relations between teaching robot's sensor and motor；
4th step：Imitate robot and this mapping relations is applied to itself control strategy, imitate the learning by imitation of teaching robot's behavior；
5th step：Appropriation is analyzed.
2. the robot learning by imitation method based on Gaussian process according to claim 1, it is characterised in that：Two parameters in the second step are optical sensor output valve and corresponding motor output valve.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510805218.9A CN106773659A (en) | 2015-11-20 | 2015-11-20 | A kind of robot learning by imitation method based on Gaussian process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510805218.9A CN106773659A (en) | 2015-11-20 | 2015-11-20 | A kind of robot learning by imitation method based on Gaussian process |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106773659A true CN106773659A (en) | 2017-05-31 |
Family
ID=58916872
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510805218.9A Pending CN106773659A (en) | 2015-11-20 | 2015-11-20 | A kind of robot learning by imitation method based on Gaussian process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106773659A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109710507A (en) * | 2017-10-26 | 2019-05-03 | 北京京东尚科信息技术有限公司 | A kind of method and apparatus of automatic test |
| CN110293560A (en) * | 2019-01-12 | 2019-10-01 | 鲁班嫡系机器人（深圳）有限公司 | Robot behavior training, planing method, device, system, storage medium and equipment |
| CN111983922A (en) * | 2020-07-13 | 2020-11-24 | 广州中国科学院先进技术研究所 | Robot demonstration teaching method based on meta-simulation learning |
-
2015
- 2015-11-20 CN CN201510805218.9A patent/CN106773659A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109710507A (en) * | 2017-10-26 | 2019-05-03 | 北京京东尚科信息技术有限公司 | A kind of method and apparatus of automatic test |
| CN109710507B (en) * | 2017-10-26 | 2022-03-04 | 北京京东尚科信息技术有限公司 | Automatic testing method and device |
| CN110293560A (en) * | 2019-01-12 | 2019-10-01 | 鲁班嫡系机器人（深圳）有限公司 | Robot behavior training, planing method, device, system, storage medium and equipment |
| CN111983922A (en) * | 2020-07-13 | 2020-11-24 | 广州中国科学院先进技术研究所 | Robot demonstration teaching method based on meta-simulation learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106773659A (en) | A kind of robot learning by imitation method based on Gaussian process | |
| CN108406767A (en) | Robot autonomous learning method towards man-machine collaboration | |
| WO2007076516A3 (en) | Brain-based device having a cerebellar model for predictive motor control | |
| CN204706208U (en) | Intelligence children education robot | |
| WO2013181109A3 (en) | Dynamical event neuron and synapse models for learning spiking neural networks | |
| CN109858574A (en) | The autonomous learning method and system of intelligent body towards man-machine coordination work | |
| Durán et al. | Rob's robot: Current and future challenges for humanoid robots | |
| CN103886367A (en) | Bionic intelligent control method | |
| Mohan et al. | How past experience, imitation and practice can be combined to swiftly learn to use novel “tools”: Insights from skill learning experiments with baby humanoids | |
| Falahi et al. | Adaptive handshaking between humans and robots, using imitation: Based on gender-detection and person recognition | |
| US20220001292A1 (en) | Programmable toy building blocks system | |
| Habib | Interdisciplinary mechatronics engineering science and the evolution of human friendly and adaptive mechatronics | |
| Moorthy et al. | Imitation based training to enhance psychomotor skills in autistic children using a snatcher robot | |
| Duarte et al. | Automatic synthesis of controllers for real robots based on preprogrammed behaviors | |
| Caci et al. | Robotics: a new tool for education of subjects with cognitive diseases | |
| Guneysu et al. | Attitudes of children towards dancing robot nao: A kindergarden observation | |
| Vignolo et al. | Using robot adaptivity to support learning in child-robot interaction | |
| Chuang et al. | Learning of composite actions and visual categories via grounded linguistic instructions: Humanoid robot simulations | |
| Christensen | Why AI Can't Be Education's Cure-All | |
| Cannone | Educational robotics in informal contexts: an experience at CoderDojo Pomezia | |
| Ramírez et al. | Object Affordances by Inferring on the Surroundings | |
| Hilleli et al. | Deep learning of robotic tasks without a simulator using strong and weak human supervision | |
| Kober et al. | Policy Search for Motor Primitives. | |
| Low et al. | Contribution Toward Future Biorobots [TC Spotlight] | |
| Jianjun et al. | The research and implementation of behavior imitation system about Nao robot based on Kinect [J] |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170531 |
|
| WD01 | Invention patent application deemed withdrawn after publication |