Search Results

Forecasting the motion of the leading vehicle is a critical task for connected autonomous vehicles as it provides an efficient way to model the leading-following vehicle behavior and analyze the interactions. In this study, a personalized time-series modeling approach for leading vehicle trajectory prediction considering different driving styles is proposed. The method enables a precise, personalized trajectory prediction for leading vehicles with limited inter-vehicle communication signals, such as vehicle speed, acceleration, space headway, and time headway of the front vehicles. Based on the learning nature of human beings that a human always tries to solve problems based on grouping and similar experience, three different driving styles are first recognized based on an unsupervised clustering with a Gaussian Mixture Model (GMM).

The detection and tracking of the damage process between surfaces in contact, together with an estimation of the remaining service life, are significant contributions to the efficient operation of hydraulic components. The commonly used approach of analyzing vibration signals in terms of spectral distributions, while being very effective, has some shortcomings. For example, the results are sensitive to both load and speed variations. The approach presented in this paper is based on the fact that the asperity distribution of surfaces in good condition have a near normal probability distribution. Deviation from this can be tracked using statistical moments. The Beta probability distribution provides a number of shapes, including normal, under the control of two positive numbers, α and β. Unlike the normal distribution, which indicates defects by kurtosis values higher than 3.0, the Beta distribution provides more flexibility.

Life Cycle Assessment (LCA) is commonly used to measure the environmental and economic impacts of engineering projects and/or products. However, there is some uncertainty associated with any LCA study. The LCA inventory analysis generally relies on imperfect data in addition to further uncertainties created by the assessment process itself. It is necessary to measure the effects that data and process uncertainty have on the LCA result and to communicate the level of uncertainty to those making decisions based on the LCA. To accomplish this, a systematic and rigorous means to assess the overall uncertainty in LCA results is required. This paper demonstrates the use of Monte Carlo Analysis to track and measure the propagation of uncertainty in LCA studies. The Monte Carlo technique basically consists of running repeated assessments using random input values chosen from a specified probable range.

Tessellation methods have been applied to characterize second phase particle fields and the degree of clustering present in AA 5754 and 5182 automotive sheet alloys. A model of damage development within these materials has been developed using a damage percolation approach based on measured particle distributions. The model accepts tessellated particle fields in order to capture the spatial distributions of particles, as well as nearest neighbour and cluster parameter data. The model demonstrates how damage initiates and percolates within particle clusters in a stable fashion for the majority of the deformation history. Macro-cracking leading to final failure occurs as a chain reaction with catastrophic void linkage triggered once linkage beyond three or more clusters of voids takes place.

The characterization of sheet metals under in-plane uniaxial bending is challenging due to the aspect ratios involved that can cause buckling. Anti-buckling plates can be employed but require compensation for contact pressure and friction effects. Recently, a novel in-plane bending fixture was developed to allow for unconstrained sample rotation that does not require an anti-buckling device. The objective of the present study is to design the sample geometry for sheared edge fracture characterization under in-plane bending along with a methodology to resolve the strains exactly at the edge. A series of virtual experiments were conducted for a 1.0 mm thick model material with different hardening rates to identify the influence of gage section length, height, and the radius of the transition region on the bend ratio and potential for buckling. Two specimen geometries are proposed with one suited for constitutive characterization and the other for sheared edge fracture.

The effects of bead surface roughness on friction, die pickup, and sheet surface damage in the drawbead test were investigated. Beads of HRC 58 hardness were prepared from centerless-ground rod by circumferential honing to 0.05 μm roughness, followed by finishing with 100, 400, or 600 grit SiC paper in the axial direction. Paraffinic base oils with viscosities of 4.5, 30, and 285 mm2/s were used neat and in conjunction with stearic acid. The effects of bead roughness depended on the nature of metal transfer, especially its distribution and firmness of attachment. The presence of a boundary additive increased, decreased, or had no effect on friction depending on the particular coating and bead finish.

Porous silicon as gas/chemical sensing material has been widely investigated in recent years. In this paper, the humidity sensing property of n-type porous silicon with ordered structure is studied for the first time. The ordered porous silicon used in this experiment has uniform pore size, pore shape and distribution. Both the membrane and closed bottom samples were studied. The resistance change of the porous silicon was measured. A 22-28% decrease of resistance was observed when relative humidity was changed from 1% to 100%. Both the response time and the recovery time were within 10 minutes, and 90% of the response can be reached in 6 minutes for the PS membrane sample. The possible sensing mechanism and future work are also discussed in this paper.

This paper presents the numerical validation of the impact response of a hot formed B-pillar component with tailored properties. A laboratory-scale B-pillar tool is considered with integral heating and cooling sections in an effort to locally control the cooling rate of an austenitized blank, thereby producing a part with tailored microstructures to potentially improve the impact response of these components. An instrumented falling-weight drop tower was used to impact the lab-scale B-pillars in a modified 3-point bend configuration to assess the difference between a component in the fully hardened (martensitic) state and a component with a tailored region (consisting of bainite and ferrite). Numerical models were developed using LS-DYNA to simulate the forming and thermal history of the part to estimate the final thickness and strain distributions as well as the predicted microstructures.

The work presented in this paper utilizes advanced FE models of the pre-bending and hydroforming process to investigate the effect of bending boost and hydroforming end-feed on the hydroformability of a tube. A model of a rotary-draw tube bender was used to simulate pre-bending of DP600 tube after which models of hydroforming of the pre-bent tube were run with various levels of end-feed. By varying bending boost from low (LB), medium (MB) and high (HB), consistent trends in the strain and thickness distribution within the pre-bent tubes were observed. Three end-feed levels were simulated and showed that an increase in end-feed improved formability during hydroforming. The sensitivity of the models to bending boost was shown.

Robust lane marking detection remains a challenge, particularly in temperate climates where markings degrade rapidly due to winter conditions and snow removal efforts. In previous work, dynamic Bayesian networks with heuristic features were used with the feature distributions trained using semi-supervised expectation maximization, which greatly reduced sensitivity to initialization. This work has been extended in three important respects. First, the tracking formulation used in previous work has been corrected to prevent false positives in situations where only poor RANSAC hypotheses were generated. Second, the null hypothesis is reformulated to guarantee that detected hypotheses satisfy a minimum likelihood. Third, the computational requirements have been greatly reduced by computing an upper bound on the marginal likelihood of all part hypotheses upon generation and rejecting parts with an upper bound less likely than the null hypothesis.

The performance characteristics of a commercial lean-NOX trap catalyst were evaluated between 200 and 500°C, using H2, CO, and a mixture of both H2 and CO as reductants before and after different high-temperature aging steps, from 600 to 750°C. Tests included NOX reduction efficiency during cycling, NOX storage capacity (NSC), oxygen storage capacity (OSC), and water-gas-shift (WGS) and NO oxidation reaction extents. The WGS reaction extent at 200 and 300°C was negatively affected by thermal degradation, but at 400 and 500°C no significant change was observed. Changes in the extent of NO oxidation did not show a consistent trend as a function of thermal degradation. The total NSC was tested at 200, 350 and 500°C. Little change was observed at 500°C with thermal degradation but a steady decrease was observed at 350°C as the thermal degradation temperature was increased.

Traditional electrode materials for lithium-ion storage cells are typically crystalline layered structures such as metal oxides, and graphitic carbons. These materials power billions of portable electronic devices in today's society. However, large-scale, high-capacity storage devices capable of powering hybrid electric vehicles (HEV″s) or their plug-in versions (PHEV's) have much more demanding requirements with respect to safety, cost, and the power they must deliver. Recently, nanostructured solid state materials, which are comprised of two more compositional or structural phases, have been found to show exciting possibilities to meet these criteria.

Safety assurance is a central concern for the development and societal acceptance of automated driving (AD) systems. Perception is a key aspect of AD that relies heavily on Machine Learning (ML). Despite the known challenges with the safety assurance of ML-based components, proposals have recently emerged for unit-level safety cases addressing these components. Unfortunately, AD safety cases express safety requirements at the system level and these efforts are missing the critical linking argument needed to integrate safety requirements at the system level with component performance requirements at the unit level. In this paper, we propose the Integration Safety Case for Perception (ISCaP), a generic template for such a linking safety argument specifically tailored for perception components. The template takes a deductive and formal approach to define strong traceability between levels.