Title :
Stable, online learning using CMACs for neuroadaptive tracking control of flexible-joint manipulators
Author :
Macnab, C.J.B. ; D´Eleuterio, G.M.T.
Author_Institution :
Inst. for Aerosp. Studies, Toronto Univ., Downsview, Ont., Canada
Abstract :
An artificial neural network is proposed for the precision control of flexible-joint robots. The training method uses backstepping in an online, direct neuroadaptive scheme in order to guarantee stability. The online weight updates include a learning term that improves performance while maintaining stability. Albus´s cerebellar model arithmetic computer algorithm is modified to work for flexible robots by utilizing radial basis functions to deal with the elasticity. The resulting hybrid network is referred to as CMAC-RBF associative memory or CRAM network. Many of the properties of the CMAC for rigid robot control are kept by using CRAM for flexible-joint robots
Keywords :
adaptive control; cerebellar model arithmetic computers; content-addressable storage; feedforward neural nets; learning (artificial intelligence); manipulators; neurocontrollers; position control; stability; CMACs; associative memory; backstepping; cerebellar model arithmetic computer algorithm; flexible-joint manipulators; neuroadaptive tracking control; online direct neuroadaptive scheme; precision control; radial basis functions; stable online learning; Adaptive control; Artificial neural networks; Backstepping; Computer networks; Digital arithmetic; Equations; Gears; Robot control; Stability; Transmission line matrix methods;
Conference_Titel :
Robotics and Automation, 1998. Proceedings. 1998 IEEE International Conference on
Conference_Location :
Leuven
Print_ISBN :
0-7803-4300-X
DOI :
10.1109/ROBOT.1998.677025
