Weighted subspace methods and spatial smoothing: analysis and comparison

The effect of using a spatially smoothed forward-backward covariance matrix on the performance of weighted eigen-based state space methods/ESPRIT, and weighted MUSIC for direction-of-arrival (DOA) estimation is analyzed. Expressions for the mean-squared error in the estimates of the signal zeros and the DOA estimates, along with some general properties of the estimates and optimal weighting matrices, are derived. A key result is that optimally weighted MUSIC and weighted state-space methods/ESPRIT have identical asymptotic performance. Moreover, by properly choosing the number of subarrays, the performance of unweighted state space methods can be significantly improved. It is also shown that the mean-squared error in the DOA estimates is independent of the exact distribution of the source amplitudes. This results in a unified framework for dealing with DOA estimation using a uniformly spaced linear sensor array and the time series frequency estimation problems.

Chance constrained uncertain classification via robust optimization

This paper studies the problem of constructing robust classifiers when the training is plagued with uncertainty. The problem is posed as a Chance-Constrained Program (CCP) which ensures that the uncertain data points are classified correctly with high probability. Unfortunately such a CCP turns out to be intractable. The key novelty is in employing Bernstein bounding schemes to relax the CCP as a convex second order cone program whose solution is guaranteed to satisfy the probabilistic constraint. Prior to this work, only the Chebyshev based relaxations were exploited in learning algorithms. Bernstein bounds employ richer partial information and hence can be far less conservative than Chebyshev bounds. Due to this efficient modeling of uncertainty, the resulting classifiers achieve higher classification margins and hence better generalization. Methodologies for classifying uncertain test data points and error measures for evaluating classifiers robust to uncertain data are discussed. Experimental results on synthetic and real-world datasets show that the proposed classifiers are better equipped to handle data uncertainty and outperform state-of-the-art in many cases.

Surface-ray tracing on hybrid surfaces of revolution for UTD mutual coupling analysis

An angle invariance property based on Hertz's principle of particle dynamics is employed to facilitate the surface-ray tracing on nondevelopable hybrid quadric surfaces of revolution (h-QUASOR's). This property, when used in conjunction with a Geodesic Constant Method, yields analytical expressions for all the ray-parameters required in the UTD formulation. Differential geometrical considerations require that some of the ray-parameters (defined heuristically in the UTD for the canonical convex surfaces) be modified before the UTD can be applied to such hybrid surfaces. Mutual coupling results for finite-dimensional slots have been presented as an example on a satellite launch vehicle modeled by general paraboloid of revolution and right circular cylinder.

Ionic conductivity, mechanical strength and Li-ion battery performance of mono-functional and bi-functional (''Janus'') ``soggy sand'' electrolytes

Influence of dispersion of uniformly sized mono-functional and bi-functional (''Janus'') particles on ionic conductivity of novel ``soggy sand'' electrolytes and its implications on mechanical strength and lithium-ion battery performance are discussed here.

A characteristic based numerical method with tracking for nonlinear wave equations

Many physical problems can be modeled by scalar, first-order, nonlinear, hyperbolic, partial differential equations (PDEs). The solutions to these PDEs often contain shock and rarefaction waves, where the solution becomes discontinuous or has a discontinuous derivative. One can encounter difficulties using traditional finite difference methods to solve these equations. In this paper, we introduce a numerical method for solving first-order scalar wave equations. The method involves solving ordinary differential equations (ODEs) to advance the solution along the characteristics and to propagate the characteristics in time. Shocks are created when characteristics cross, and the shocks are then propagated by applying analytical jump conditions. New characteristics are inserted in spreading rarefaction fans. New characteristics are also inserted when values on adjacent characteristics lie on opposite sides of an inflection point of a nonconvex flux function, Solutions along characteristics are propagated using a standard fourth-order Runge-Kutta ODE solver. Shocks waves are kept perfectly sharp. In addition, shock locations and velocities are determined without analyzing smeared profiles or taking numerical derivatives. In order to test the numerical method, we study analytically a particular class of nonlinear hyperbolic PDEs, deriving closed form solutions for certain special initial data. We also find bounded, smooth, self-similar solutions using group theoretic methods. The numerical method is validated against these analytical results. In addition, we compare the errors in our method with those using the Lax-Wendroff method for both convex and nonconvex flux functions. Finally, we apply the method to solve a PDE with a convex flux function describing the development of a thin liquid film on a horizontally rotating disk and a PDE with a nonconvex flux function, arising in a problem concerning flow in an underground reservoir.

Construction of solid model from measured point data

This paper describes an algorithm for constructing the solid model (boundary representation) from pout data measured from the faces of the object. The poznt data is assumed to be clustered for each face. This algorithm does not require any compuiier model of the part to exist and does not require any topological infarmation about the part to be input by the user. The property that a convex solid can be constructed uniquely from geometric input alone is utilized in the current work. Any object can be represented a5 a combznatzon of convex solids. The proposed algorithm attempts to construct convex polyhedra from the given input. The polyhedra so obtained are then checked against the input data for containment and those polyhedra, that satisfy this check, are combined (using boolean union operation) to realise the solid model. Results of implementation are presented.

Local polynomial convexity of the union of two totally-real surfaces at their intersection

We consider the following question: Let S (1) and S (2) be two smooth, totally-real surfaces in C-2 that contain the origin. If the union of their tangent planes is locally polynomially convex at the origin, then is S-1 boolean OR S-2 locally polynomially convex at the origin? If T (0) S (1) a (c) T (0) S (2) = {0}, then it is a folk result that the answer is yes. We discuss an obstruction to the presumed proof, and provide a different approach. When dim(R)(T0S1 boolean AND T0S2) = 1, we present a geometric condition under which no consistent answer to the above question exists. We then discuss conditions under which we can expect local polynomial convexity.

Variable Sparsity Kernel Learning

This paper(1) presents novel algorithms and applications for a particular class of mixed-norm regularization based Multiple Kernel Learning (MKL) formulations. The formulations assume that the given kernels are grouped and employ l(1) norm regularization for promoting sparsity within RKHS norms of each group and l(s), s >= 2 norm regularization for promoting non-sparse combinations across groups. Various sparsity levels in combining the kernels can be achieved by varying the grouping of kernels-hence we name the formulations as Variable Sparsity Kernel Learning (VSKL) formulations. While previous attempts have a non-convex formulation, here we present a convex formulation which admits efficient Mirror-Descent (MD) based solving techniques. The proposed MD based algorithm optimizes over product of simplices and has a computational complexity of O (m(2)n(tot) log n(max)/epsilon(2)) where m is no. training data points, n(max), n(tot) are the maximum no. kernels in any group, total no. kernels respectively and epsilon is the error in approximating the objective. A detailed proof of convergence of the algorithm is also presented. Experimental results show that the VSKL formulations are well-suited for multi-modal learning tasks like object categorization. Results also show that the MD based algorithm outperforms state-of-the-art MKL solvers in terms of computational efficiency.

Ultrafast Microwave-Assisted Route to Surfactant-Free Ultrafine Pt Nanoparticles on Graphene: Synergistic Co-reduction Mechanism and High Catalytic Activity

We demonstrate an ultrafast method for the formation of, graphene supported Pt catalysts by the co-reduction of graphene oxide and Pt salt using ethylene glycol under microwave irradiation conditions. Detailed analysis of the mechanism of formation of the hybrids indicates a synergistic co-reduction mechanism whereby the presence of the Pt ions leads to a faster reduction of GO and the presence of the defect sites on the reduced GO serves as anchor points for the heterogeneous nucleation of Pt. The resulting hybrid consists of ultrafine nanoparticles of Pt uniformly distributed on the reduced GO susbtrate. We have shown that the hybrid exhibits good catalytic activity for methanol oxidation and hydrogen conversion reactions. The mechanism is general and applicable for the synthesis of other multifunctional hybrids based on graphene.

Shape optimization of RC flexural members

A natural velocity field method for shape optimization of reinforced concrete (RC) flexural members has been demonstrated. The possibility of shape optimization by modifying the shape of an initially rectangular section, in addition to variation of breadth and depth along the length, has been explored. Necessary shape changes have been computed using the sequential quadratic programming (SQP) technique. Genetic algorithm (Goldberg and Samtani 1986) has been used to optimize the diameter and number of main reinforcement bars. A limit-state design approach has been adopted for the nonprismatic RC sections. Such relevant issues as formulation of optimization problem, finite-element modeling, and solution procedure have been described. Three design examples-a simply supported beam, a cantilever beam, and a two-span continuous beam, all under uniformly distributed loads-have been optimized. The results show a significant savings (40-56%) in material and cost and also result in aesthetically pleasing structures. This procedure will lead to considerable cost saving, particularly in cases of mass-produced precast members and a heavy cast-in-place member such as a bridge girder.

Effect of porosity on the indentation behaviour of nanoporous alumina films

Anodising aluminium in an acid electrolyte results in a porous alumina film. The pores are uniformly distributed and their structure can be controlled by varying the anodising parameters. In this study, the pore structure of the anodic alumina films is varied systematically by varying the anodising time and a subsequent pore widening chemical etch. The indentation behaviour of the resulting film is studied using a depth sensing nanoindenter. The hardness of the films was found to be decreasing with decreasing solid area fraction. Understanding the deformation behaviour of nanoporous alumina would help tailoring the mechanical properties by tuning the geometry.

Turbulent near-wake development behind a flat plate

An asymptotic analysis of the two-dimensional turbulent near-wake flow behind a Rat plate with sharp trailing edge has been formulated, The feature that the near-wake, which is dominated by the mixing of the oncoming turbulent boundary layers retains, to a large extent, the memory of the turbulent structure of the upstream boundary layer has been exploited to develop the analysis. This analysis leads to two regions of the near-wake flow (the inner near-wake and the outer near-wake) for which the governing equations are derived. The matching conditions among these regions lead to a logarithmic variation in the normal direction in the overlapping region surrounding the inner near-wake. These features are validated by the available experimental data. Similarity solutions for the velocity distribution (which satisfy the required matching conditions) in the inner near-wake and outer near-wake regions have been obtained by making the appropriate eddy-viscosity assumptions, Uniformly valid solutions for velocity distribution have been constructed for the near-wake. The solutions show good agreement with available experimental data. (C) Elsevier, Paris.

Profile-Based Load Scheduling in Wireless Energy Harvesting Sensors for Data Rate Maximization

In this paper, power management algorithms for energy harvesting sensors (EHS) that operate purely based on energy harvested from the environment are proposed. To maintain energy neutrality, EHS nodes schedule their utilization of the harvested power so as to save/draw energy into/from an inefficient battery during peak/low energy harvesting periods, respectively. Under this constraint, one of the key system design goals is to transmit as much data as possible given the energy harvesting profile. For implementational simplicity, it is assumed that the EHS transmits at a constant data rate with power control, when the channel is sufficiently good. By converting the data rate maximization problem into a convex optimization problem, the optimal load scheduling (power management) algorithm that maximizes the average data rate subject to energy neutrality is derived. Also, the energy storage requirements on the battery for implementing the proposed algorithm are calculated. Further, robust schemes that account for the insufficiency of battery storage capacity, or errors in the prediction of the harvested power are proposed. The superior performance of the proposed algorithms over conventional scheduling schemes are demonstrated through computations using numerical data from solar energy harvesting databases.

Robust Relay Precoder Design for MIMO-Relay Networks

In this paper, we consider a robust design of MIMO-relay precoder and receive filter for the destination nodes in a non-regenerative multiple-input multiple-output (MIMO) relay network. The network consists of multiple source-destination node pairs assisted by a single MIMO-relay node. The source and destination nodes are single antenna nodes, whereas the MIMO-relay node has multiple transmit and multiple receive antennas. The channel state information (CSI) available at the MIMO-relay node for precoding purpose is assumed to be imperfect. We assume that the norms of errors in CSI are upper-bounded, and the MIMO-relay node knows these bounds. We consider the robust design of the MIMO-relay precoder and receive filter based on the minimization of the total MIMO-relay transmit power with constraints on the mean square error (MSE) at the destination nodes. We show that this design problem can be solved by solving an alternating sequence of minimization and worst-case analysis problems. The minimization problem is formulated as a convex optimization problem that can be solved efficiently using interior-point methods. The worst-case analysis problem can be solved analytically using an approximation for the MSEs at the destination nodes. We demonstrate the robust performance of the proposed design through simulations.

Robust Formulations for Handling Uncertainty in Kernel Matrices

We study the problem of uncertainty in the entries of the Kernel matrix, arising in SVM formulation. Using Chance Constraint Programming and a novel large deviation inequality we derive a formulation which is robust to such noise. The resulting formulation applies when the noise is Gaussian, or has finite support. The formulation in general is non-convex, but in several cases of interest it reduces to a convex program. The problem of uncertainty in kernel matrix is motivated from the real world problem of classifying proteins when the structures are provided with some uncertainty. The formulation derived here naturally incorporates such uncertainty in a principled manner leading to significant improvements over the state of the art. 1.