Artif Intell Rev (2007) 27:113–130
DOI 10.1007/s10462-008-9087-0
Recurrent neural robot controllers: feedback mechanisms
for identifying environmental motion dynamics
Stephen Paul McKibbin · Bala Amavasai ·
Arul N. Selvan · Fabio Caparrelli · W. A. F. W. Othman
Published online: 28 October 2008
© Springer Science+Business Media B.V. 2008
Abstract In this paper a series of recurrent controllers for mobile robots have been devel-
oped. The system combines the iterative learning capability of neural controllers and the
optimisation ability of particle swarms. In particular, three controllers have been developed:
an Exo-sensing, an Ego-sensing and a Composite controller which is the hybrid of the latter
two. The task for each controller is to learn to follow a moving target and identify its trajec-
tory using only local information. We show how the learned behaviours of each architecture
rely on different sensory representations, although good results are obtained in all cases.
Keywords Dynamic Neural Networks · Particle Swarm Optimisation · Sensory-Motor
Coordination
1 Introduction
In nature, animals use the sensory information that is available to them in order to benefit
them in completing a task. In some cases this information is readily available to them at
instantaneously and can be acted upon immediately and in other cases that same information
S. P. McKibbin (B ) · B. Amavasai · A. N. Selvan · F. Caparrelli · W.A. F. W. Othman
Microsystems & Machine Vision Laboratory, Materials & Engineering Research Institute,
Sheffield Hallam University, Sheffield S1 1WB, UK
e-mail: S.Mckibbin@shu.ac.uk; Stephen.p.mckibbin@student.shu.ac.uk
B. Amavasai
Mechanistic Research, Procter & Gamble, 460 Basingstoke Road, Reading RG2 0QE, UK
e-mail: amavasai.b@pg.com
A. N. Selvan
e-mail: A.N.Selvan@shu.ac.uk
F. Caparrelli
e-mail: F.Caparrelli@shu.ac.uk
W. A. F. W. Othman
e-mail: Wan.A.Othman@student.shu.ac.uk
123