Robot control using high dimensional neural networks
暂无分享,去创建一个
In this paper, we propose a position control scheme for actual robot system using high dimensional neural networks. Complex-valued neural network and quaternion neural network learn inverse kinematics of the robot systems. Control objectives are two dimensional SCARA robot and three dimensional robot. Tip of these robots are controlled by the high dimensional neural networks. Some results by an actual robot system are shown to confirm feasibility of these high dimensional neural networks as robot controllers.
[1] Rui J. P. de Figueiredo,et al. Learning rules for neuro-controller via simultaneous perturbation , 1997, IEEE Trans. Neural Networks.
[2] John H. Conway,et al. On Quaternions and Octonions , 2003 .
[3] G. Dixon,et al. On quaternions and octonions: Their geometry, arithmetic, and symmetry , 2004 .
[4] Nobuyuki Matsui,et al. Quaternionic Neural Networks: Fundamental Properties and Applications , 2009 .
[5] T. Ogawa. Complex-Valued Neural Network and Inverse Problems , 2009 .