Behind-armor debris that results from tungsten rods penetrating armor steel at 2 km/s was studied by analysis of recovered fragments. Fragment recovery was by means of particle board. Individual fragments were analyzed by x-ray tomography, which provides information for fragment identification, mass, shape, and penetration down to masses of a few milligrams. The experiments were complemented by AUTODYN and EPIC calculations. Fragments were steel or tungsten generated from the channel or from the breakout through the target rear surface. Channel fragment motions were well described by Tate theory. Breakout fragments had velocities from the projectile remnant to the channel velocity, apparently depending on where in the projectile a fragment originated. The fragment size distribution was extremely broad and did not correlate well with simple uniform-fragment-size models.

A comparison of techniques for obtaining projectile velocity history on a two-stage launcher and discuss gun code accuracy vis-a-vis pressure gauges and the new photonic Doppler velocimetry (PDV) technique is presented. The PDV technique itself is described in a companion paper. The PDV records were differentiated to compute acceleration and, hence, base pressure. Two acceleration episodes are revealed in the data. Base pressure values were compared with measurements from stationary pressure gauges and with predictions of a standard two-stage gun code. The agreement with the pressure gages was satisfactory. Code predictions did not account for the two acceleration stages. However, for the main acceleration episode, the predicted base pressure is in good agreement with the smoothed pressure computed from the PDV record. Both the gauge records and PDV contain short-time pressure spikes which are apparently real. Therefore, use of computed base pressure for projectile design may lead to failures if the projectile is vulnerable to pressure spikes.

Bar impact tests, using the techniques described elsewhere in this symposium, were used to measure compressive and tensile strengths of borosilicate glass, soda lime glass, and a glass ceramic. The glass ceramic was 25% crystalline spinel, furnished by Corning Inc. There are two measures of compressive strength: the peak stress that can be transmitted in unconfined compression, and the "steady state" strength. For borosilicate glass and soda lime glass, these values were similar, being about 1.8 and 1.5 GPa, respectively. The glass ceramic (25% spinel) was almost 50% stronger. Tensile failure in the glass and glass ceramic takes places via surface flaws, and thus tensile strength is an extrinsic, as opposed to intrinsic property.

High-velocity impact onto a layered glass target produces a very extensive damage pattern exhibiting many distinct morphologies. High-speed photography reveals failure waves and cracks that move at acoustic velocities. These prompt features evolve into a complex final damage pattern that includes needle fragments around the penetration cavity, radial cracks at mid distance, and dicing cracks near the edges.

In any high-energy pulsed power experiment, the metallic conductors are expected to heat up significantly due to resistive losses. In the pulsed case, the effects of local heat transfer are decreased due to the limited thermal diffusion time, so the process is considered to be adiabatic rather than isothermal. Previous results indicate that the high-temperature mechanical properties of metallic conductors significantly depend on the rapidity and duration of heat deposition. With this in mind, it is important to understand the mechanical properties of metals heated rapidly so that the correct mechanical properties are considered when designing high energy experiments. An expanding ring experiment has been performed at the Institute for Advanced Technology (IAT) to test such mechanical properties. The experiment uses a primary coil powered by a near-critically damped RC circuit to induce a current pulse in a thin specimen ring that expands and fragments due to electromagnetic forces. So that the heating time is minimized, an inductive heating source has been developed to rapidly heat the specimen ring. Temperatures as high as the material's melting temperature can be reached within a few milliseconds, prior to the application of electromagnetic expansion forces. The source employs a pumped LC tank circuit with a resonant frequency of roughly 25 kHz to induce a current in the ring. The current in the primary and secondary coils are measured using Pearson and Rogowski coils. A high-speed infrared camera is used to measure the temperature of the ring specimen during heating.

Experiments have been conducted with 6.25 mm diameter tungsten rods striking concrete at 2.2 km/s. Three concretes were used-one was 2.35 g/cm(3) and the other two were 2.27 g/cm(3). The erosion rates were measured to be T/Delta L=2.4-3.1 depending on the density of the concrete. This is greater than the hydrodynamic value, which shows that the strength of the penetrator is affecting the penetration. The cratering efficiency was computed (which included surface spall) and was found to be commensurate with the strength of the concrete, 28-34 MPa. CTH calculations were conducted using the brittle fracture kinetics (BFK) and Holmquist-Johnson-Cook (HJC) material models for concrete. Density in the calculations was 2.25 g/cm(3). It was not possible to match erosion rates at 2.2 km/s, which were too high in the calculations. Also, computed crater volumes were much too small, mainly due to spall in the experiments that was not shown in the computations. Another significant inaccuracy of the calculations was the damage extent, which became unrealistically widespread as time increased in the BFK model. (C) 2008 Elsevier Ltd. All rights reserved.

A fluid dynamic formulation of speech generation may lead to an improved understanding of the physics of speech production. Unlike more traditional linear acoustic methods of speech synthesis, this alternate approach aims to capture more of the relevant physics by numerically solving a form of the Reynolds-Averaged Navier-Stokes equations describing fluid motion. Though computationally intensive, the method is not limited by assumptions of linearity and plane wave propagation inherent in linear acoustic analysis. Numerical simulations of flows in stylized vocal tract shapes, as weil as measurements on physical flows are described. Special attention is given to fricative generation, since the physiological understanding, and subsequent synthesis, of these sounds stands to gain the most from this approach.

Author's preface.1 Introduction2 Preliminaries2.1 The physics of speech production2.2 The source-filter model2.3 Information-bearing features of the speech signal2.4 Time-frequency representations2.5 Classifications of acoustic patterns in speech2.6 Temporal invariance and stationarity2.7 Taxonomy of linguistic structure3 Mathematical models of linguistic structure3.1 Probabilistic functions of a discrete Markov process3.2 Formal grammars and abstract automata4 Syntactic analysis4.1 Deterministic parsing algorithms4.2 Probabilistic parsing algorithms4.3 Parsing natural language5 Grammatical inference5.1 Exact inference and Gold's theorem5.2 Baum's algorithm for regular grammars5.3 Event counting in parse trees5.4 Baker's algorithm for context-free grammars6 Information-theoretic analysis of speech communication6.1 The Miller et al. experiments6.2 Entropy of an information source6.3 Recognition error rates and entropy7 Automatic speech recognition and constructive theories of language7.1 Integrated architectures7.2 Modular architectures7.3 Parameter estimation from fluent speech7.4 System performance7.5 Other speech technologies8 Automatic speech understanding and semantics8.1 Transcription and comprehension8.2 Limited domain semantics8.3 The semantics of natural language8.4 System architectures8.5 Human and machine performance9 Theories of mind and language9.1 The challenge of automatic natural language understanding9.2 Metaphors for mind9.3 The artificial intelligence program10 A speculation on the prospects for a science of the mind10.1 The parable of the thermos bottle: measurements and symbols10.2 The four questions of science10.3 A constructive theory of the mind10.4 The problem of consciousness10.5 The role of sensorimotor function, associative memory and reinforcement learning in automatic acquisition of spoken language by an autonomous robot10.6 Final thoughts: predicting the course of discovery

This chapter contains sections titled: Transcription and comprehension Limited domain semantics The semantics of natural language System architectures Human and machine performance

This chapter contains sections titled: Integrated architectures Modular architectures Parameter estimation from fluent speech System performance Other speech technologies

This paper presents a novel fused hidden Markov model (fused HMM) for integrating tightly coupled time series, such as audio and visual features of speech. In this model, the time series are first modeled by two conventional HMMs separately. The resulting HMMs are then fused together using a probabilistic fusion model, which is optimal according to the maximum entropy principle and a maximum mutual information criterion. Simulations and bimodal speaker verification experiments show that the proposed model can significantly reduce the recognition errors in noiseless or noisy environments.

World EnglishesVolume 22, Issue 3 p. 227-232 Ethical implications of an experiment in artificial intelligence Stephen E. Levinson, Stephen E. Levinson 1 University of Illinois at Urbana-Champaign, USA sel@ifp.uiuc.eduSearch for more papers by this author 1 Stephen E. Levinson, Stephen E. Levinson 1 University of Illinois at Urbana-Champaign, USA sel@ifp.uiuc.eduSearch for more papers by this author 1 First published: 19 August 2003 https://doi.org/10.1111/1467-971X.00292AboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat Volume22, Issue3August 2003Pages 227-232 RelatedInformation

An important research problem for solid armature railguns is understanding the sudden transition from low-voltage sliding electrical contact between the rails and armature to high-voltage contact involving a plasma arc. Although an armature can fail in many ways and thereby "transition," two distinct transition mechanisms currently limit the performance of solid armature electromagnetic (EM) launchers. One mechanism, broadly termed "wear-induced transition," results from uneven or excessive loss of material from the contact interface. The other mechanism is associated with a rapid reduction in driving current (negative dI/dt). A series of experiments have shown the waveform-induced transition mechanism is independent of the state of wear of an armature. This paper explores this mechanism, which we call "electrodynamic transition," and seek to explain the cause of transition as an electromechanical instability that disrupts the liquid film interface between armature and rail. Using 3-D finite element analyses (FEA), we observed the development of localized forces at the edges of the armature as the driving current drops rapidly. This behavior, we believe, may cause molten material to be ejected from the armature-rail contact region.