Robustness of representations in multilayer feedforward neural networks

Recent research has shown that multilayer feedforward networks with sigmoidal activation functions are universal approximators, and that this holds for more general activations as well. The mathematical underpinning for these results has been various: Kolmogorov's resolution of Hilbert's thirteenth problem; the Stone-Weierstrass theorem; approximation of Fourier and Radon integral representations; and convergence of probability measures. This paper

Chaotic exploration and learning of locomotion behaviors

We present a general and fully dynamic neural system, which exploits intrinsic chaotic dynamics, for the real-time goal-directed exploration and learning of the possible locomotion patterns of an articulated robot of an arbitrary morphology in an unknown environment. The controller is modeled as a network of neural oscillators that are initially coupled only through physical embodiment, and goal-directed exploration of coordinated motor patterns is achieved by chaotic search using adaptive bifurcation. The phase space of the indirectly coupled neural-body-environment system contains multiple transient or permanent self-organized dynamics, each of which is a candidate for a locomotion behavior. The adaptive bifurcation enables the system orbit to wander through various phase-coordinated states, using its intrinsic chaotic dynamics as a driving force, and stabilizes on to one of the states matching the given goal criteria. In order to improve the sustainability of useful transient patterns, sensory homeostasis has been introduced, which results in an increased diversity of motor outputs, thus achieving multiscale exploration. A rhythmic pattern discovered by this process is memorized and sustained by changing the wiring between initially disconnected oscillators using an adaptive synchronization method. Our results show that the novel neurorobotic system is able to create and learn multiple locomotion behaviors for a wide range of body configurations and physical environments and can readapt in realtime after sustaining damage.     

Preparation of AZ91D Slurries for Semi-Solid Forming Using Al8 ( Mn, Fe) 5 Precipitates

During the solidification of the AZ91D-alloys, the Al8(Mn, Fe)5 phase is generally precipitated in the melt in advance of the precipitation of the primary α-Mg. The basic principle for manufacturing AZ91D-alloy slurries for semi solid forming is to use the Al8(Mn, Fe)5 precipitates as the heterogeneous nucleation sites for primary α-Mg phases. Microscopic analysis for the location of the Al8(Mn, Fe)5 precipitate explains that the Al8(Mn, Fe)5 precipitate is the effective heterogeneous nucleation site for the primary α-Mg phase. It was also observed that increase of the Mn content in the melt and the cooling rate below to the solid/liquid two-phase region resulted in smaller and more globular primary α-Mg due to the increase of heterogeneous nucleation sites. It was found that the average α-Mg diameter grew as a function of t0.278, where t is the holding time at the solid/liquid two-phase region. This would be attributed to the Ostwald type ripening and coalescence between primary α-Mg phases. The cooling rate below to the solid/liquid two-phase region, Mn content in AZ91D alloy, and the holding time and temperature affected on the quality of slurry.     

3D television: Developing a multi‐modal multi‐viewer TV system of the future

Abstract— Under the European Union funded Advanced Three‐dimensional Television System Technologies (ATTEST) project, De Montfort University (DMU) is developing a 3D display system targeted specifically at domestic television applications. This system uses a modified transmissive LCD panel together with novel backlighting and optics to project multiple viewing regions, or exit pupils, into the viewing space. These exit pupils are located in space using a head tracker. The display accommodates multiple viewers simultaneously and imposes no physical constraints, such as special eyewear. Viewers may move freely over a “typical” room‐sized area. The design of the backlighting facilitates many other display regimes beyond the “standard” 3DTV mode in which each viewer sees the same image pair.     

Counting Fluorescent Dye Molecules on DNA Origami by Means of Photon Statistics

Obtaining quantitative information about molecular assemblies with high spatial and temporal resolution is a challenging task in fluorescence microscopy. Single‐molecule techniques build on the ability to count molecules one by one. Here, a method is presented that extends recent approaches to analyze the statistics of coincidently emitted photons to enable reliable counting of molecules in the range of 1–20. This method does not require photochemistry such as blinking or bleaching. DNA origami structures are labeled with up to 36 dye molecules as a new evaluation tool to characterize this counting by a photon statistics approach. Labeled DNA origami has a well‐defined labeling stoichiometry and ensures equal brightness for all dyes incorporated. Bias and precision of the estimating algorithm are determined, along with the minimal acquisition time required for robust estimation. Complexes containing up to 18 molecules can be investigated non‐invasively within 150 ms. The method might become a quantifying add‐on for confocal microscopes and could be especially powerful in combination with STED/RESOLFT‐type microscopy.     

Terrain classification and identification of tree stems using ground‐based LiDAR

To operate autonomously in forested terrain, unmanned ground vehicles must be able to identify the load‐bearing surface of the terrain (i.e., the ground) and obstacles in the environment. To travel long distances, they must be able to track their position even when the forest canopy obstructs GPS signals, e.g., by tracking progress relative to tree stems. This paper presents a novel, robust approach for modeling the ground plane and tree stems in forests from a single viewpoint using a lightweight LiDAR scanner. Ground plane identification is implemented using a two‐stage approach. The first stage, a local height‐based filter, discards most nonground points. The second stage, based on a support vector machine classifier, identifies which of the remaining points belong to the ground. Main tree stems are modeled as cylinders or cones to estimate the diameter 130 cm above the ground plane. To fit these models, candidate main stem data are selected by finding points approximately 130 cm above the ground. These points are clustered into separate point clouds for each stem. Cylinders and cones are fit to each point cloud, and heuristic filters identify which fits correspond to tree stems. Experimental results from five forested environments demonstrate the effectiveness of this approach. For ground plane estimation, the overall classification accuracy was 86.28% with a mean error for the ground height of approximately 4.7 cm. For stem estimation, up to 50% of the main stems were accurately modeled using cones, with a root mean square diameter error of 13.2 cm.© 2012 Wiley Periodicals, Inc.