Adaptive dynamic balance of a biped robot using neural networks

Author

Kun, Andrija ; Miller, W. Thomas, III

Author_Institution

Dept. of Electr. & Comput. Eng., New Hampshire Univ., Durham, NH, USA

Volume

1

fYear

1996

fDate

22-28 Apr 1996

Firstpage

240

Abstract

An adaptive dynamic balance scheme was implemented and tested on an experimental biped. The control scheme used pre-planned but adaptive motion sequences. CMAC neural networks were responsible for the adaptive control of side-to-side and front-to-back balance, as well as for maintaining good foot contact. Qualitative and quantitative test results show that the biped performance improved with neural network training. The biped is able to start and stop on demand, and to walk with continuous motion on flat surfaces at a rate of up to 100 steps per minute, with up to 6 cm long step

Keywords

adaptive control; cerebellar model arithmetic computers; legged locomotion; mobile robots; motion control; neurocontrollers; robot dynamics; CMAC neural networks; adaptive control; adaptive dynamic balance; adaptive motion sequences; biped robot; continuous motion; flat surfaces; foot contact; front-to-back balance; neural network training; side-to-side balance; Adaptive control; Foot; Kinematics; Legged locomotion; Motion control; Neural networks; Programmable control; Robot sensing systems; Service robots; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Robotics and Automation, 1996. Proceedings., 1996 IEEE International Conference on

Conference_Location

Minneapolis, MN

ISSN

1050-4729

Print_ISBN

0-7803-2988-0

Type

conf

DOI

10.1109/ROBOT.1996.503784

Filename

503784