Use this URL to cite or link to this record in EThOS:
Title: Moving softly : the role of morphological computation in the generation of intelligent behaviour
Author: Johnson, Chris
Awarding Body: University of Sussex
Current Institution: University of Sussex
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
The central theme of this thesis is 'morphological computation', in particular as it pertains to the generation of purposeful and intelligent behaviour. The first part of the thesis covers experiments in simulations, aimed at exploring a previously unasked question: can genuinely computational soft-body reservoirs play a central role in generating behaviour which has previously been described as 'minimally cognitive'? We conclude in the affirmative, but also note that it remains unclear whether or how they can also exceed such abilities. The main part of this work has been presented at ALIFE 14 and ECAL 15, and published in the Artificial Life journal. In part two, the focus moves to consideration of how morphological computation in soft bodies may collude with the action of nervous systems in the production of adaptive intelligent behaviour. The massive and hypnotic complexity of brains in present day species, and lasting traditions in the design and control of animals' mechanical analogies (robots), still lead many to perceive bodies in the neuromuscular species as being in service to the goals of their controllers (brains), but in fact the two are of a piece, and work together to the same ends. We hypothesise that this would have been more apparent than now at a time closer to the evolutionary introduction of neuromuscular systems, and introduce our new software framework for the evolution of virtual neuromuscular swimmers, which may be used to study artificial parallels to those distant origins. We begin with a demonstration of the technical innovations which we required in order to simulate soft-bodied swimmers in two-dimensional particle-based fluids, and then proceed to a description of our new concept for mapping arbitrary neural networks to arbitrary body morphologies.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Q0334 Artificial intelligence ; QP0351 Neurophysiology and neuropsychology ; TJ0210.2 Mechanical devices and figures. Automata. Ingenious mechanisms. Robots (General)