Imitative motion generation for humanoid robots based on the motion knowledge learning and reuse

Okuzawa, Yuki; Kato, Shōhei; Kanoh, Masayoshi; Ide, Hidenori

doi:10.1109/icsmc.2009.5346686

Cited by 4 publications

(3 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Okuzawa et al [27] introduced an imitation based motion generation model for humanoid robots. Recognition of instructed movement primitives is accomplished by using HMMs.…”

Section: Imitation Learningmentioning

confidence: 99%

Simple and complex behavior learning using behavior hidden Markov model and CobART

2013

View full text Add to dashboard Cite

Date:I hereby declare that all information in this document has been obtained and presented in accordance with academic rules and ethical conduct. I also declare that, as required by these rules and conduct, I have fully cited and referenced all material and results that are not original to this work. In this thesis, behavior learning and generation models are proposed for simple and complex behaviors of robots using unsupervised learning methods. Simple behaviors are modeled by simple-behavior learning model (SBLM) and complex behaviors are modeled by complex-behavior learning model (CBLM) which uses previously learned simple or complex behaviors. Both models have common phases named behavior categorization, behavior modeling, and behavior generation. Sensory data are categorized using correlation based adaptive resonance theory network that generates motion primitives corresponding to robot's base abilities in the categorization phase. In the modeling phase, Behavior-HMM, a modified version of hidden Markov model, is used to model the relationships among the motion primitives in a finite state stochastic network. In addition, a motion generator which is an artificial neural network is trained for each motion primitive to learn essential robot motor commands. In the generation phase, desired task is presented as a target observation and the model generates corresponding motion primitive sequence. Then, these motion primitives are executed successively by the motion generators which are specifically trained for the corresponding motion primitives.The models are not proposed for one specific behavior, but are intended to be bases for all behaviors. CBLM enhances learning capabilities by integrating previously learned behaviors hierarchically. Hence, new behaviors can take advantage of already discovered behaviors. The proposed models are tested on a robot simulator and the experiments showed that simple and complex-behavior learning models can generate requested behaviors effectively. Bu tez kapsamında, basit ve karmaşık robot davranışlarını gözetmen ihtiyacı duymayan yöntemlerle ögrenen ve tekrarlayabilen davranış modelleri tasarlandı. Basit davranışlar basit-davranış ögrenme modeli (SBLM) ile karmaşık davranışlar ise daha önce ögrenilen basit ve karmaşık davranışları kullanabilen karmaşık-davranış ögrenme modeli (CBLM) ile modellendi. Her iki model de davranış sınıflan-dırma, davranış modelleme ve davranış üretme aşamalarından oluşmaktadır. Sınıflandırma aşamasında algılayıcılardan elde edilen veriler robotun temel yeteneklerine karşılık gelen temel hareketleri elde etmek amacıyla ilinti temelli uyarlanır rezonans kuramı kullanılarak sınıflandırılır. Modelleme aşama-sında, saklı Markov modelinin degiştirilmiş birşekli olan davranış-HMM kullanılarak temel hareketler arasındaki ilişki sonlu durumlu olasılıksal ag biçiminde modellenir. Davranış modeline ilaveten robot motor komutlarını ögrenmek amacıyla her bir temel davranış için yapay sinir aglarını kullanan bir hareket üretici egitilir. Davranış üretme aşamasın...

show abstract

“…Okuzawa et al [27] introduced an imitation based motion generation model for humanoid robots. Recognition of instructed movement primitives is accomplished by using HMMs.…”

Section: Imitation Learningmentioning

confidence: 99%

Simple and complex behavior learning using behavior hidden Markov model and CobART

2013

View full text Add to dashboard Cite

show abstract

“…We proposed imitative motion generation system using the discrete HMM [7]. In this paper, we used the continuous HMM in recognizing approach.…”

Section: Introductionmentioning

confidence: 99%

Acquisition and modification of motion knowledge using continuous HMMs for motion imitation of humanoids

Okuzawa

Kato

Kanoh

2009

2009 International Symposium on Micro-NanoMechatronics and Human Science

Self Cite

View full text Add to dashboard Cite

A knowledge-based approach to imitation learning of motion generation for humanoid robots and an imitative motion generation system based on motion knowledge learning and reuse are described. The system has three parts: recognizing, learning, and modifying parts. The first part recognizes an instructed motion distinguishing it from the motion knowledge database by the continuous hidden markov model. When the motion is recognized as being unfamiliar, the second part learns it using locally weighted regression and acquires a knowledge of the motion. When a robot recognizes the instructed motion as familiar or judges that its acquired knowledge is applicable to the motion generation, the third part imitates the instructed motion by modifying a learned motion. This paper reports some performance results: the motion imitation of several radio gymnastics motions.

show abstract

“…Imitated human motions are associated with existing robot motions whereby the robot motions are assigned to certain tasks or labels. To avoid duplications in the robot motions stored in the library, researchers used variations of Hidden Markov Models (HMM) to differentiate between similar and different motions [Kulić et al, 2008, Okuzawa et al, 2009, Calinon et al, 2010. Others used methods like Principal Component Analysis [Motomura et al, 2009, Tran et al, 2010 or expectation-maximization clustering algorithms [Sung et al, 2012].…”

Section: Comparisons Of Motion Trajectoriesmentioning

confidence: 99%

Autonomous animation of humanoid robots

Tay¹

View full text Add to dashboard Cite

from NTU. I am also grateful for the administrative support given by Christina Contreras and Chris Hertz at CMU, Agnes Tan, Eng Cheng Lim and Meow Chng Soh at NTU, which has enabled me to focus on my research. I would like to express my gratitude and appreciation to Changjiu Zhou, Gerald Seet and Nadine Aubry for their advice regarding the NTU-CMU Dual Ph.D. programme. I am also grateful to enjoy the support of my friends in the NTU-CMU Dual Ph.D. programme: Chun Fan Goh, Guo Zhan Lum, Juanjuan Zhang and Wei Sin Ang, and the Singaporeans in CMU: Ying Ying Wu, Wee Liat Ong and Matthew Lee. I am very grateful to the Economic Development Board of Singapore that graciously funds my graduate studies at NTU and CMU.

show abstract

Imitative motion generation for humanoid robots based on the motion knowledge learning and reuse

Cited by 4 publications

References 8 publications

Simple and complex behavior learning using behavior hidden Markov model and CobART

Simple and complex behavior learning using behavior hidden Markov model and CobART

Acquisition and modification of motion knowledge using continuous HMMs for motion imitation of humanoids

Autonomous animation of humanoid robots

Contact Info

Product

Resources

About