Evolutionary robotics: the Sussex approach

We give an overview of evolutionary robotics research at Sussex over the last five years. We explain and justify our distinctive approaches to (artificial) evolution, and to the nature of robot control systems that are evolved. Results are presented from research with evolved controllers for autonomous mobile robots, simulated robots, co-evolved animats, real robots with software controllers, and a real robot with a controller directly evolved in hardware.     

Knowledge-based vision and simple visual machines

The vast majority of work in machine vision emphasizes the representation of perceived objects and events: it is these internal representations that incorporate the 'knowledge' in knowledge-based vision or form the 'models' in model-based vision. In this paper, we discuss simple machine vision systems developed by artificial evolution rather than traditional engineering design techniques, and note that the task of identifying internal representations within such systems is made difficult by the lack of an operational definition of representation at the causal mechanistic level. Consequently, we question the nature and indeed the existence of representations posited to be used within natural vision systems (i.e. animals). We conclude that representations argued for on a priori grounds by external observers of a particular vision system may well be illusory, and are at best place-holders for yet-to-be-identified causal mechanistic interactions. That is, applying the knowledge-based vision approach in the understanding of evolved systems (machines or animals) may well lead to theories and models that are internally consistent, computationally plausible, and entirely wrong.     

Gamma interferon production by cytotoxic T lymphocytes is required for resolution of Chlamydia trachomatis infection

In this study, we used mice in which the gene for gamma interferon (IFN-gamma) has been disrupted (IFN-gamma-/- mice) to study the role of this cytokine in the resolution of Chlamydia trachomatis infection. We show that IFN-gamma-/- mice are impaired in the ability to clear infection with C. trachomatis compared to IFN-gamma+/+ control mice. Activated CD8(+) cytotoxic T lymphocytes (CTL) secrete IFN-gamma in response to intracellular infection, and we have shown previously that a Chlamydia-specific CTL line can reduce C. trachomatis infection when adoptively transferred into infected mice. In the present study, we found that when these IFN-gamma+/+ CTL lines are transferred into Chlamydia-infected IFN-gamma-/- mice, the transferred CTL cannot overcome the immune defect seen in the IFN-gamma-/- mice. We also show that Chlamydia-specific CTL can be cultured from IFN-gamma-deficient mice infected with C. trachomatis; however, the adoptive transfer of IFN-gamma-/- CTL into infected IFN-gamma+/+ mice does not reduce the level of infection. These results suggest that IFN-gamma production by CTL is not sufficient to overcome the defect that IFN-gamma-/- mice have in the resolution of Chlamydia infection, yet IFN-gamma production by CTL is required for the protective effect seen upon adoptive transfer of CTL into IFN-gamma+/+ mice.     

Singular-perturbation of the time-optimal soft-constrainedcheap-control problem

The solution of a time-optimal soft-constrained control problem with linear dynamics is considered. The cost function has no penalty on the integral of the state. The solution is formulated in terms of the controllability Grammian and is obtained as the solution of a system of linear algebraic equations and a nonlinear scalar algebraic equation. As the state approaches the origin, or equivalently, as the control becomes cheap, the optimal final time becomes small. This introduces a highly degenerate hierarchy of amplitude scales. An approach based solely on expanding the controllability Grammian is developed to obtain an asymptotic solution of the problem without resorting to boundary-layer theory     

Interfacial debonding in laminated titanium matrix composites

The results of an experimental program in which multiaxial loads were applied to [0₄] and [±45]_s silicon carbide/titanium (SiC/Ti) tubes are reviewed showing that stress coupling, matrix viscoplasticity (including room temperature creep) and fiber/matrix interfacial damage all contribute to nonlinear response and permanent strains in titanium matrix composites (TMC). A micromechanical model that explicitly considers the aforementioned phenomena is presented herein. The model assumes a periodic microstructure and uses finite elements to analyze a representative volume element. The composite is assumed to be in a state of generalized plane strain making it possible to discretize only a generic transverse plane while still being able to apply three-dimensional loading through appropriate boundary conditions. The response of laminated composites is predicted by incorporating the micromechanical results into nonlinear lamination theory. Predictions are presented to show the influence of the model parameters on the effective composite response of unidirectional [0₄] and angle-ply [±45]_s TMC laminates.     

In memoriam Jerold L. Swedlow

In memoriam Jerold L. Swedlow     

Error Rate Analysis for Coded Multicarrier Systems Over Quasi-Static Fading Channels

Several standards such as IEEE 802.11a/g, IEEE 802.16, and the European Computer Manufacturers Association (ECMA) multiband orthogonal frequency division multiplexing (MB-OFDM) for high data-rate ultra-wideband employ bit-interleaved convolutionally coded multicarrier modulation over quasi-static fading channels. Motivated by the lack of appropriate error rate analysis techniques for this popular type of system and channel model, we present two novel analytical methods for bit error rate (BER) estimation of coded multicarrier systems operating over frequency-selective quasi-static channels with nonideal interleaving. In the first method, the approximate performance of the system is calculated for each realization of the channel, which is suitable for obtaining the outage BER performance (a common performance measure for, e.g., MB-OFDM systems). The second method assumes Rayleigh distributed frequency-domain subcarrier channel gains and knowledge of their correlation matrix, and can be used to directly obtain the average BER performance. Both methods are applicable to convolutionally coded interleaved multicarrier systems employing quadrature amplitude modulation, and are also able to account for narrowband interference (modeled as a sum of tone interferers). To illustrate the application of the proposed analysis, both methods are used to study the performance of a tone-interference-impaired MB-OFDM system.