Abstract The control of a robot arm is a particularly difficult task due to the influence of non-linearities, disturbances, gravity and load effects. In the present work a computer model of a 3-joint robot arm which is conceived for orange picking is used as a working example of such a kind of robots. It will be shown how Fuzzy Logic Controllers (FLCs) can suitably perform the control task. A procedure to obtain and to optimize FLCs is also described. Then, with the aim of realizing an adaptive control system capable of compensate for the unmodeled nonlinearities, disturbances and time variance effects of the robot, an Artificial Neural Network (ANN) is adopted into a feed-forward scheme using the Radial Basis Function network (RBFN) model.
[1]
George N. Saridis,et al.
Compensation of Industrial Manipulator Dynamics
,
1989,
Int. J. Robotics Res..
[2]
Roberto Horowitz,et al.
Learning Control of Robot Manipulators
,
1993
.
[3]
R.T. Newton,et al.
Neural network control of a space manipulator
,
1993,
IEEE Control Systems.
[4]
Gang Feng.
Improved Tracking Control for Robots Using Neural Networks
,
1993,
1993 American Control Conference.
[5]
Marcelo H. Ang,et al.
Performance of a neuro-model-based robot controller: adaptability and noise rejection
,
1992
.