Search results

This paper aims to present a novel scheme of multiple vanishing points (VPs) estimation and corresponding lanes identification.

The scheme proposed here includes two main stages: VPs estimation and lane identification. VPs estimation based on vanishing direction hypothesis and Bayesian posterior probability estimation in the image Hough space is a foremost contribution, and then VPs are estimated through an optimal objective function. In lane identification stage, the selected linear samples supervised by estimated VPs are clustered based on the gradient direction of linear features to separate lanes, and finally all the lanes are identified through an identification function.

The scheme and algorithms are tested on real data sets collected from an intelligent vehicle. It is more efficient and more accurate than recent similar methods for structured road, and especially multiple VPs identification and estimation of branch road can be achieved and lanes of branch road can be identified for complex scenarios based on Bayesian posterior probability verification framework. Experimental results demonstrate VPs, and lanes are practical for challenging structured and semi-structured complex road scenarios.

A Bayesian posterior probability verification framework is proposed to estimate multiple VPs and corresponding lanes for road scene understanding of structured or semi-structured road monocular images on intelligent vehicles.

This paper aims to study distance determination in vehicles, which could allow an in-car system to provide feedback and alert drivers, by either prompting the driver to take preventative action or prepare the vehicle’s safety systems for an imminent collision. The success of a new system's deploying allows drivers to oppose the huge number of accidents and the material losses and costs associated with car accidents.

In this context, this paper presents estimation distance between camera and frontal vehicles based on camera calibration by combining three main steps: vanishing point extraction, lanes detection and vehicles detection in the field of 3 D real scene. This algorithm was implemented in MATLAB, and it was applied on scenes containing several vehicles in highway urban area. The method starts with the camera calibration. Then, the distance information can be calculated.

Based on experiment performance, this new method achieves robustness especially for detecting and estimating distances for multiple vehicles in a single scene. Also, this method demonstrates a higher accuracy detection rate of 0.869 in an execution time of 2.382 ms.

The novelty of the proposed method consists firstly on the use of an adaptive segmentation to reject the false points of interests. Secondly, the use of vanishing point has reduced the cost of using memory. Indeed, the part of the image above the vanishing point will not be processed and therefore will be deleted. The last benefit is the application of this new method on structured roads.

The purpose of this paper is to develop a numerical (finite-difference) model exploring phase and group velocities of SH-wave propagation in initially stressed transversely isotropic poroelastic multi-layered composite structures and initially stressed viscoelastic-dry-sandy multi-layered composite structures in two distinct cases.

With the aid of relevant constitutive relations, the non-vanishing equations of motions for the propagation SH-wave in the considered composite structures have been derived. Haskell matrix method and finite-difference scheme are adopted to deduce velocity equation for both the cases. Stability analysis for the adopted finite-difference scheme has been carried out and the expressions for phase as well as group velocity in terms of dispersion-parameter and stability-ratio have been deduced.

Velocity equations are derived for the propagation of SH-wave in both the composite structures. The obtained results are matched with the classical results for the case of double and triple-layered composite structure along with comparative analysis. Stability analysis have been carried out to develop expressions of phase as well as group velocity in terms of dispersion-parameter and stability-ratio. The effect of wavenumber, dispersion parameter along with initial-stress, porosity, sandiness, viscoelasticity, stability ratio, associated with the said composite structures on phase, damped and group velocities of SH-wave has been unveiled.

To the best of authors’ knowledge, numerical modelling and analysis of propagation characteristics of SH-wave in multi-layered initially stressed composite structures composed of transversely isotropic poroelastic materials and viscoelastic-dry-sandy materials remain unattempted inspite of its importance and relevance in many branches of science and engineering.

Solves a coupled electromagnetic‐mechanical problem ‐ that of a cantilevered conductive plate in crossed steady and time‐varying magnetic fields ‐ by using a semi‐analytical method and an eddy current model. Assumes that the magnetic flux variation induces two electromotive forces to the plate; one due to the time‐varying magnetic field and the other to the plate movement in the steady magnetic field. Considers two equivalent LR circuits and notes that the superposition of their currents yields the total circulating current in the plate. Couples this electromagnetic model to the one‐dimensional beam bending model, adopted for the structural analysis of the problem and derives the system of the coupled electromagnetic‐mechanical equations. Calculates analytically some of the parameters involved in these equations so that the numerical computation of the remaining unknowns is very rapid. Concludes that the results are in agreement with the experiment and with results obtained by numerical methods treating, in three dimensions, the electromagnetic aspects of the problem. Notes that the fully numerical methods are very much CPU time consuming.

The numerical computation of magnetization processes in moving and rotating assemblies requires the usage of vector hysteresis models. A commonly used model is the so-called Mayergoyz vector Preisach model, which applies the scalar Preisach model into multiple angles of the halfspace. The usage of several scalar models, which are optionally weighted differently, enables the description of isotropic as well as anisotropic materials. The flexibility is achieved, however, at the cost of multiple scalar model evaluations. For solely isotropic materials, two vector Preisach models, based on an extra rotational operator, might offer a lightweight alternative in terms of evaluation cost. The study aims at comparing the three mentioned models with respect to computational efficiency and practical applicability.

The three mentioned vector Preisach models are compared with respect to their computational costs and their representation of magnetic polarization curves measured by a vector vibrating sample magnetometer.

The results prove the applicability of all three models to practical scenarios and show the higher efficiency of the vector models based on rotational operators in terms of computational time.

Although the two vector Preisach models, based on an extra rotational operator, have been proposed in 2012 and 2015, their practical application and inversion has not been tested yet. This paper not only shows the usability of these particular vector Preisach models but also proves the efficiency of a special stageless evaluation approach that was proposed in a former contribution.

Obtaining the maximum possible flow rates that can be induced by free convection in open‐ended vertical eccentric annuli under fundamental thermal boundary conditions of the fourth kind (heating or cooling one of the annulus walls with a uniform heat flux while keeping the other wall at ambient temperature). Obtaining the maximum possible flow rates that can be induced by free convection in open‐ended vertical eccentric annuli under fundamental thermal boundary conditions of the fourth kind (heating or cooling one of the annulus walls with a uniform heat flux while keeping the other wall at ambient temperature).

The fully‐developed laminar free convection momentum equation has been solved numerically using an analytical solution of the governing energy equation.

Results are presented to show the effect of the annulus radius ratio and the dimensionless eccentricity on the induced flow rate, the total heat absorbed by the fluid, and the fully developed Nusselt numbers on the two boundaries of the annulus for a fluid of Prandtl number 0.7.

Applications of the obtained results can be of value in the heat‐exchanger industry, in cooling of underground electric cables, and in cooling small vertical electric motors and generators.

The paper presents a solution that is not available in the literature for the problem of fully developed free convection in open‐ended vertical eccentric annular channels under thermal boundary conditions of the fourth kind. Also presents the maximum possible induced flow rates, the total heat absorbed by the fluid, and the Nusselt numbers on the two boundaries of the annulus. The effects of N and E (the radius ratio and eccentricity, respectively) on these results are presented. Such results are very much needed for design purposes of heat transfer equipment.

Robot automatic grasping has important application value in industrial applications. Recent works have explored on the performance of deep learning for robotic grasp detection. They usually use oriented anchor boxes (OABs) as detection prior and achieve better performance than previous works. However, the parameters of their loss belong to different coordinates, this may affect the regression accuracy. This paper aims to propose an oriented regression loss to solve the problem of inconsistency among the loss parameters.

In the oriented loss, the center coordinates errors between the ground truth grasp rectangle and the predicted grasp rectangle rotate to the vertical and horizontal of the OAB. And then the direction error is used as an orientation factor, combining with the errors of the rotated center coordinates, width and height of the predicted grasp rectangle.

The proposed oriented regression loss is evaluated on the YOLO-v3 framework to the grasp detection task. It yields state-of-the-art performance with an accuracy of 98.8% and a speed of 71 frames per second with GTX 1080Ti on Cornell datasets.

This paper proposes an oriented loss to improve the regression accuracy of deep learning for grasp detection. The authors apply the proposed deep grasp network to the visual servo intelligent crane. The experimental result indicates that the approach is accurate and robust enough for real-time grasping applications.

The purpose of this paper is to consider a detailed investigation of transversal magnetic (TM) nonlinear magnetooptical integrated optical sensor. The sensitivities of two sensors are presented. The first sensor composed of a dielectric thin film surrounded by a lossless, nonmagnetic, isotropic cladding exhibiting a local Kerr‐like dielectric nonlinearity, and a magnetic substrate chosen to be an iron garnet. The second sensor is formed by exchanging the cladding and the substrate media of the first sensor. The homogenous sensitivities of both sensors are calculated as a function of the waveguide thickness and the effective refractive index. The effect of nonlinearity on the sensitivities for both sensors is investigated.

The homogenous sensitivities of both sensors are calculated as a function of the waveguide thickness and the effective refractive index. The effect of nonlinearity on the sensitivities for both sensors is investigated. Numerical calculations are performed using the Maple program.

It was found that the sensitivity for the first sensor sensitivity increases with nonlinearity. While the sensitivity for the second sensor is hardly affected by the change of nonlinearity. It was also found that the thickness of the guiding layer is a critical parameter for the sensitivity of the optical sensor with the optimum thickness being just above cut‐off in case of the first structure and at the cut‐off in the case of the second structure.

A detailed investigation of TM nonlinear magnetooptical integrated optical sensor is considered. The two proposed structures are used to investigate the parameters to get the optimal sensitivity, which is an important issue is the sensor design.

The purpose of this paper is to suggest a new approach to the numerical simulation of shallow‐water flows both in plane domains and on the sphere.

The approach involves the technique of splitting of the model operator by geometric coordinates and by physical processes. Specially chosen temporal and spatial approximations result in one‐dimensional finite difference schemes that conserve the mass and the total energy. Therefore, the mass and the total energy of the whole two‐dimensional split scheme are kept constant too.

Explicit expressions for the schemes of arbitrary approximation orders in space are given. The schemes are shown to be mass‐ and energy‐conserving, and hence absolutely stable because the square root of the total energy is the norm of the solution. The schemes of the first four approximation orders are then tested by simulating nonlinear solitary waves generated by a model topography. In the analysis, the primary attention is given to the study of the time‐space structure of the numerical solutions.

The approach can be used for the numerical simulation of shallow‐water flows in domains of both Cartesian and spherical geometries, providing the solution adequate from the physical and mathematical standpoints in the sense of keeping its mass and total energy constant even when fully discrete shallow‐water models are applied.