Skill learning based catching motion control

Creative Commons License

Cimen G., Kavafoglu Z., Kavafoglu E., Çapın T. K., GÜRÇAY H.

COMPUTER ANIMATION AND VIRTUAL WORLDS, cilt.26, sa.3-4, ss.217-225, 2015 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 26 Sayı: 3-4
  • Basım Tarihi: 2015
  • Doi Numarası: 10.1002/cav.1659
  • Dergi Adı: COMPUTER ANIMATION AND VIRTUAL WORLDS
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.217-225
  • Anahtar Kelimeler: ball catching simulation, physics-based character animation, reinforcement learning
  • TED Üniversitesi Adresli: Evet

Özet

Learned biomechanical strategies prepare the human body in kinematics and kinetics terms during interception tasks, such as throwing and catching, in real world. Based on this, we present a real-time physics-based approach that generates natural and physically plausible motions for a highly complex task ball catching. We showed that ball catching behavior could be achieved with the proper combination of rather simple motor skills, such as standing, walking, and reaching. The character learns some policies to know how and when to react by using reinforcement learning in order to use time accurately. We demonstrate the effectiveness of our method with some of the catching animation results in different catching strategies. Copyright (c) 2015John Wiley & Sons, Ltd.