Solution for Ill-Posed Inverse Kinematics of Robot Arm by Network Inversion

Table 1

Network parameters.

Input neurons		3
Hidden neurons		10
Output neurons		2
Training rate ε t		0.01
Inverse estimation rate ε e		0.01
Training error to be attained		<0.001
Estimation error to be attained		<0.0001
Joint angles for training data (${}^{°}$)	$\theta$1	0, 15,$\dots$, 90
	$\theta$2	45, 60,$\dots$, 315
	$\theta$3	45, 60,$\dots$, 315
Arm length l (cm)		30