Design and optimization of a fuzzy-neural hybrid controller for an artificial muscle robotic arm using genetic algorithms

Humanoids are increasingly used in the service sectors around the world to work with, or assist humans. However current humanoid designs place limitations on direct engagement with the human in terms of safety and usability. In this paper, we present an approach for the control of hybrid, high-speed and safe human-robot interaction systems with highly non-linear dynamic behavior. The proposed approach comprises the three soft computing techniques, namely back propagation neural network, fuzzy and genetic algorithms. This open-loop controller was applied to a Bridgestone Hybrid Robot Arm (BHRA). BHRA has three electric motors and four artificial muscles, arranged in an agonist/antagonist, and opposing pair configuration, that drive the five-degrees of freedom of the robot arm. The behaviors of the artificial muscles are observed under the effects of the links driven by the electric motors and it is shown that the proposed biologically-plausible controller could produce more accurate trajectories at higher speeds when compared to conventional PID and stand alone or combined versions of Neural Network and Fuzzy controllers.

[1]  Yannis A. Phillis,et al.  Fuzzy assessment of machine flexibility , 1998 .

[2]  Daniel W. Repperger,et al.  Fuzzy PD+I learning control for a pneumatic muscle , 2003, The 12th IEEE International Conference on Fuzzy Systems, 2003. FUZZ '03..

[3]  Fumiya Iida,et al.  Toward a human-like biped robot with compliant legs , 2009, Robotics Auton. Syst..

[4]  Kyuwan Choi,et al.  Prediction of Arm Trajectory from the Neural Activities of the Primary Motor Cortex Using a Modular Artificial Neural Network Model , 2007, ICONIP.

[5]  M. Paradiso,et al.  Neuroscience: Exploring the Brain , 1996 .

[6]  J.H. Lilly.  Adaptive tracking for pneumatic muscle actuators in bicep and tricep configurations , 2003, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[7]  Lois C. McInnes,et al.  Using the GA and TAO toolkits for solving large-scale optimization problems on parallel computers , 2007, TOMS.

[8]  N. Tsagarakis,et al.  A 7 DOF pneumatic muscle actuator (pMA) powered exoskeleton , 1999, 8th IEEE International Workshop on Robot and Human Interaction. RO-MAN '99 (Cat. No.99TH8483).

[9]  D. Mitchell Wilkes,et al.  ISAC: Foundations in Human-Humanoid Interaction , 2000, IEEE Intell. Syst..

[10]  Liang Yang,et al.  Sliding mode tracking for pneumatic muscle actuators in opposing pair configuration , 2005, IEEE Transactions on Control Systems Technology.

[11]  M. Paradiso,et al.  Neuroscience: Exploring the brain, 3rd ed. , 2007 .

[12]  Kazuhiko Kawamura,et al.  A frequency modeling method of rubbertuators for control application in an IMA framework , 2001, Proceedings of the 2001 American Control Conference. (Cat. No.01CH37148).

[13]  C. S. George Lee,et al.  Neural fuzzy systems: a neuro-fuzzy synergism to intelligent systems , 1996 .

[14]  Kazuhiko Kawamura,et al.  Dynamic pneumatic actuator model for a model-based torque controller , 2003, Proceedings 2003 IEEE International Symposium on Computational Intelligence in Robotics and Automation. Computational Intelligence in Robotics and Automation for the New Millennium (Cat. No.03EX694).

[15]  K. Inoue,et al.  Defining a neural network controller structure for a rubbertuator robot , 2000, Neural Networks.

[16]  K. Kawamura,et al.  Application of a hybrid controller with non-contact impedance to a humanoid robot , 2008, 2008 International Workshop on Variable Structure Systems.

[17]  A. P. Georgopoulos,et al.  Primate motor cortex and free arm movements to visual targets in three- dimensional space. II. Coding of the direction of movement by a neuronal population , 1988, The Journal of neuroscience : the official journal of the Society for Neuroscience.

[18]  Yasuo Konishi,et al.  Nonlinear optimal predictive control of rubber artificial muscle , 1995, Other Conferences.

[19]  Kyuwan Choi,et al.  Prediction of arm trajectory from the neural activities of the primary motor cortex with modular connectionist architecture , 2009, Neural Networks.