From Models of Cognition to Robot
Control and Back Using Spiking Neural
Networks
Stefan Iacob , Johan Kwisthout , and Serge Thill
(B)
Donders Institute for Brain, Cognition and Behaviour,
Radboud University Nijmegen, 6525 Nijmegen, HR, The Netherlands
s.iacob@student.ru.nl, {j.kwisthout,s.thill}@donders.ru.nl
Abstract. With the recent advent of neuromorphic hardware there has
been a corresponding rise in interest in spiking neural network models
for the control of real-world artificial agents such as robots. Although
models of cognitive mechanisms instantiated in spiking neural networks
are nothing new, very few of them are translated onto real robot platforms. In this paper, we attempt such a translation: we implement an
existing, biologically plausible model of reaching (the REACH model)
demonstrated in 2D simulation on a UR5e robot arm. We are interested
in particular in how well such a translation works since this has implications for similar exercises with a vast library of existing models of
cognition. In this particular case, after extensions to operations in 3D
and for the particular hardware used, we do find that the model is able
to learn on the real platform as it did in the original simulation, albeit
without reaching the same levels of performance.
Keywords: Spiking neural networks · Robot control · Bio-plausible
models
1 Introduction
1.1 Biologically Inspired Artificial Cognition and Spiking Neurons
Although artificial cognitive systems and, more specifically, robotics can be
seen as engineering disciplines, there has traditionally been a strong interest
in biological approaches to designing such systems [28]. Work has, over the past
decades, spanned simple demonstrations (in terms of robot agent and environment; often virtual) that were based on hypothesised biological mechanisms (see
for example early implementations of internal simulation mechanisms [19,31]
based on Hesslow’s simulation hypothesis [15,16]) to biologically inspired control for autonomous vehicles [8].
A long-going debate in this context concerns the appropriate level at which
to base such biological inspiration, with options ranging from the purely
behavioural to the physiologically detailed. Several factors can play a role in
c
Springer Nature Switzerland AG 2020
V. Vouloutsi et al. (Eds.): Living Machines 2020, LNAI 12413, pp. 176–191, 2020.
https://doi.org/10.1007/978-3-030-64313-3_18
Control and Back Using Spiking Neural
Networks
Stefan Iacob , Johan Kwisthout , and Serge Thill
(B)
Donders Institute for Brain, Cognition and Behaviour,
Radboud University Nijmegen, 6525 Nijmegen, HR, The Netherlands
s.iacob@student.ru.nl, {j.kwisthout,s.thill}@donders.ru.nl
Abstract. With the recent advent of neuromorphic hardware there has
been a corresponding rise in interest in spiking neural network models
for the control of real-world artificial agents such as robots. Although
models of cognitive mechanisms instantiated in spiking neural networks
are nothing new, very few of them are translated onto real robot platforms. In this paper, we attempt such a translation: we implement an
existing, biologically plausible model of reaching (the REACH model)
demonstrated in 2D simulation on a UR5e robot arm. We are interested
in particular in how well such a translation works since this has implications for similar exercises with a vast library of existing models of
cognition. In this particular case, after extensions to operations in 3D
and for the particular hardware used, we do find that the model is able
to learn on the real platform as it did in the original simulation, albeit
without reaching the same levels of performance.
Keywords: Spiking neural networks · Robot control · Bio-plausible
models
1 Introduction
1.1 Biologically Inspired Artificial Cognition and Spiking Neurons
Although artificial cognitive systems and, more specifically, robotics can be
seen as engineering disciplines, there has traditionally been a strong interest
in biological approaches to designing such systems [28]. Work has, over the past
decades, spanned simple demonstrations (in terms of robot agent and environment; often virtual) that were based on hypothesised biological mechanisms (see
for example early implementations of internal simulation mechanisms [19,31]
based on Hesslow’s simulation hypothesis [15,16]) to biologically inspired control for autonomous vehicles [8].
A long-going debate in this context concerns the appropriate level at which
to base such biological inspiration, with options ranging from the purely
behavioural to the physiologically detailed. Several factors can play a role in
c
Springer Nature Switzerland AG 2020
V. Vouloutsi et al. (Eds.): Living Machines 2020, LNAI 12413, pp. 176–191, 2020.
https://doi.org/10.1007/978-3-030-64313-3_18