IEEE Transactions on Systems, Man and Cybernetics
Kinematic networks distributed model for representing and regularizing motor redundancy
Biological Cybernetics
Autonomous Robots
Forward models for physiological motor control
Neural Networks - 1996 Special issue: four major hypotheses in neuroscience
Multiple paired forward and inverse models for motor control
Neural Networks - Special issue on neural control and robotics: biology and technology
Using Humanoid Robots to Study Human Behavior
IEEE Intelligent Systems
Reaching with multi-referential dynamical systems
Autonomous Robots
Learning and generalization of motor skills by learning from demonstration
ICRA'09 Proceedings of the 2009 IEEE international conference on Robotics and Automation
ICRA'09 Proceedings of the 2009 IEEE international conference on Robotics and Automation
The iCub humanoid robot: an open platform for research in embodied cognition
PerMIS '08 Proceedings of the 8th Workshop on Performance Metrics for Intelligent Systems
Teaching humanoids to imitate 'shapes' of movements
ICANN'10 Proceedings of the 20th international conference on Artificial neural networks: Part II
Teaching a humanoid robot to draw `Shapes'
Autonomous Robots
Hi-index | 0.00 |
This paper addresses the problem of planning the movement of highly redundant humanoid robots based on non-linear attractor dynamics, where the attractor landscape is obtained by combining multiple force fields in different reference systems. The computational process of relaxation in the attractor landscape is similar to coordinating the movements of a puppet by means of attached strings, the strings in our case being the virtual force fields generated by the intended/attended goal and the other task dependent combinations of constraints involved in the execution of the task. Hence the name PMP (Passive Motion Paradigm) was given to the computational model. The method does not require explicit kinematic inversion and the computational mechanism does not crash near kinematic singularities or when the robot is asked to achieve a final pose that is outside its intrinsic workspace: what happens, in this case, is the gentle degradation of performance that characterizes humans in the same situations. Further, the measure of inconsistency in the relaxation in such cases can be directly used to trigger higher level reasoning in terms of breaking the goal into a sequence of subgoals directed towards searching and perhaps using tools to realize the otherwise unrealizable goal. The basic PMP model has been further expanded in the present paper by means of (1) a non-linear dynamical timing mechanism that provides terminal attractor properties to the relaxation process and (2) branching units that allow to `compose' complex PMP-networks to coordinate multiple kinematic chains in a complex structure, including manipulated tools. A preliminary evaluation of the approach has been carried out with the 53 degrees of freedom humanoid robot iCub, with particular reference to trajectory formation and bimanual/whole upper body coordination under the presence of different structural and task specific constraints.