Malachy Eaton

Further explorations in evolutionary humanoid robotics

The field of evolutionary humanoid robotics is a branch of evolutionary robotics specifically dealing with the application of evolutionary principles to humanoid robot design. Previous studies demonstrated the possible future potential of this approach by evolving walking behaviors for simulated humanoid robots with up to 20 degrees of freedom. In this paper we examine further the evolutionary process by looking at the changes in diversity over time. We then investigate the effect of the immobilization of an individual joint or joints in the robot. The latter study may be of potential future use in prosthetic design. We also explore the possibility of the evolution of humanoid robots which can cope with different environmental conditions. These include reduced ground friction (ice) and modified gravitation (moon walking). We present initial results on the implementation of our simulated humanoid robots in hardware using the Bioloid robotic platform, using a model of this robot in order to evolve the desired motion patterns, for subsequent transfer to the real robot. We finish the article with a summary and brief discussion of future work.

Toward a benchmarking framework for research into bio-inspired hardware-software artefacts

In this paper, we suggest that one of the more crucial tasks currently facing researchers into the field of autonomous mobile robotics is the provision of a common task, or set of tasks, as a means of evaluating different approaches to robot design and architecture, and the generation of a common set of experimental frameworks to facilitate these different approaches. This paper stars with a brief introduction to the field, and behavior-based control in particular. We then discuss the issue of animal versus robot behavior, and focus on simulated experimentation versus embodied robotics. Finally, we move to the feasibility of evaluating and benchmarking different architectures, with the aim of producing mobile robots of continuously higher utility, with specific reference to our current four-layered robot control architecture.

Evolutionary humanoid robotics: past, present and future

Evolutionary robotics is a methodology for the creation of autonomous robots using evolutionary principles. Humanoid robotics is concerned specifically with autonomous robots that are human-like in that they mimic the body or aspects of the sensory, processing and/or motor functions of humans to a greater or lesser degree. We investigate how these twin strands of advanced research in the field of autonomous mobile robotics have progressed over the last decade or so, and their current recent convergence in the new field of evolutionary humanoid robotics. We describe our current work in the evolution of controllers for bipedal locomotion in a simulated humanoid robot using an accurate physics simulator, and briefly discuss the effects of changes in robot mobility and of environmental changes. We then describe our current work in the implementation of these simulated robots using the Bioloid robot platform. We conclude with a look at possible visions for the future.

An Approach to the Synthesis of Humanoid Robot Dance Using Non-interactive Evolutionary Techniques

After bipedal locomotion, dance is one of the most commonly studied behaviours for researchers seeking to replicate human-like motion in humanoid robots. Many of the methods employed involve direct interaction with, or imitation of, human participant(s). For example, the generation of dance movements using interactive evolutionary computation (IEC) involves the replacement of an objective fitness function with the subjective evaluations of human observer(s). In this paper we present an alternative approach to the synthesis of humanoid robot dance using non-interactive evolutionary computation (non-IEC) methods. We propose a novel fitness function for the evolution of robotic dance, and we present initial results of the application of this evolutionary process to the generation of dance patterns for the 18-DOF Bioloid humanoid robot. We conclude that even without the presence of a human or humans in the evolutionary loop, it is possible to produce surprisingly lifelike and novel dances using this approach.

Evolutionary control of bipedal locomotion in a high degree-of-freedom humanoid robot: first steps

This article describes a methodology, together with an associated series of experiments employing this methodology, for the evolution of walking behavior in a simulated humanoid robot with up to 20 degrees of freedom. The robots evolved in this study learn to walk smoothly in an upright or near-upright position and demonstrate a variety of different locomotive behaviors, including “skating,” “limping,” and walking in a manner curiously reminiscent of a mildly or heavily intoxicated person. A previous study demonstrated the possible potential utility of this approach while evolving controllers based on simulated humanoid robots with a restricted range of movements. Although walking behaviors were developed, these were slow and relied on the robot walking in an excessively stooped position, similar to the gait of an infirm elderly person. This article extends the previous work to a robot with many degrees of freedom, up to 20 in total (arms, elbows, legs, hips, knees, etc.), and demonstrates the automatic evolution of fully upright bipedal locomotion in a humanoid robot using an accurate physics simulator.

Evolving Smart Initial Layouts for Force-Directed Graph Drawing

We propose a genetic algorithm (GA) for solving the maximization version of the Optimal Linear Arrangement problem and we also demonstrate how solutions found by it can be used for constructing smart initial layouts for force-directed graph drawing. Effectively, we show that our GA can be used as a first step in force-directed graph drawing for achieving more aesthetically pleasing graph layouts at the end. We present experimental results which show that the initial layouts based on the solutions of our GA reduce the number of edge crossings in force-directed graph layouts.

Artificial life and embodied robotics: current issues and future challenges

In this article we explore some of the issues currently facing researchers in the interface between the twin fields of Artificial Life and Robotics, and the challenges and potential synergy of these two areas in the creation of future robotic life forms. There are three strands of research which we feel will be of key importance in the possible development of future embodied artificial life forms. These are the areas of evolutionary robotics and evolutionary humanoid robotics in particular, probabilistic robotics for deliberation, and robot benchmarking with associated metrics and standards. We briefly explore each of these areas in turn, focusing on our current research in each field and what we see as the potential issues and challenges for the future.

Pursuit-evasion using evolutionary algorithms in an immersive three-dimensional environment

I n view of the biological prevalence of pursuit-evasion conies6 they provide a useful test-bed for research into novel bio-inspired computing and control systems. I n this paper we investigate the evolution of pursuit-evasion strategies in a virtual-reality environment created using the Unreal World Editor. The Unreal World Editor (UnrealED), original& designed for use with thep opular 3 0 game Unreal, is an easily available editor which can be used for the creation and modification of a wide variety of immersive environments I n this paper we model buildings on the University of Limerick campus, and their associated exteriors This paper makes use of an extension tot he original Unreal game engine called a mutator, more specifically the Gamebots mutator designed and released by the University of Southern California’s Information Sciences Institute. This atension allows characters in the game to be controlled via network sockets connected to other programs. The game feeds sensory information to the character over the network connection. Based on this information, the client program can decide whata etions the being should take and issues commands back over the network to the game, in order to control the actions of the entie. The client program incorporates a genetic algorithm to control the two individuals involved

A Global Representation Scheme for Genetic Algorithms

Modelling the behaviour of genetic algorithms has concentrated on Markov chain analysis. However, Markov chains yield little insight into the dynamics of the underlying mechanics and processes. Thus, a framework and methodology for global modelling and visualisation of genetic algorithms is described, using tools from the field of Information Theory. Using Principal Component Analysis (PCA) based on the Karhunen-Loeve transform, a generation (instance of a population) is transformed into a compact low dimensional eigenspace representation. A pattern vector (set of weights) is calculated for each population of strings, by projecting it into the eigenspace. A 3D manifold or global signature is derived from the set of computed pattern vectors.

A GA-Inspired Approach to the Reduction of Edge Crossings in Force-Directed Layouts

We report on our findings using a genetic algorithm (GA) as a preprocessing step for force-directed graph drawings to find a smart initial vertex layout (instead of a random initial layout) to decrease the number of edge crossings in the graph. We demonstrate that the initial layouts found by our GA improve the chances of finding better results in terms of the number of edge crossings, especially for sparse graphs and star-shaped graphs. In particular we demonstrate a reduction in edge-crossings for the class of star-shaped graphs by using our GA over random vertex placement in the order of 3:1.