Search Results

Autonomous vehicles (AVs) have already driven millions of miles on public roads, but even the simplest scenarios have not been certified for safety. Current methodologies for the verification of AV’s decision and control systems attempt to divorce the lower level, short-term trajectory planning and trajectory tracking functions from the behavioral rules-based framework that governs mid-term actions. Such analysis is typically predicated on the discretization of the state space and has several limitations. First, it requires that a conservative buffer be added around obstacles such that many feasible plans are classified as unsafe. Second, the discretized controllers modeled in this analysis require several refinement steps before being implementable on an actual AV, and typically do not allow the specification of comfort-related properties on the trajectories. Consumer-ready AVs use motion planning algorithms that generate smooth trajectories.

The stoichiometric direct injection gasoline engines have higher torque performance than the port injection engines, as the volumetric efficiency can be increased due to the cooling effects of charging air by the fuel evaporation in the cylinder. They need only 3-way catalyst, leading to the cost down. However there exists the injection timing (region) that increased volumetric efficiency does not lead to higher torque. In order to investigate the phenomena, the in-cylinder mixture formation process has been analyzed by the LIF and the CFD techniques. As the results, it has been revealed that the phenomena are caused by the inhomogeneous mixture distribution before the ignition timing.

A reduction in brain disorders owing to traumatic brain injury (TBI) caused by head impacts in traffic accidents is needed. However, the details of the injury mechanism still remain unclear. In past analyses, brain parenchyma of a head finite element (FE) model has generally been modeled using simple isotropic viscoelastic materials. For further understanding of TBI mechanism, in this study we developed a new constitutive model that describes most of the mechanical properties in brain parenchyma such as anisotropy, strain rate dependency, and the characteristic features of the unloading process. Validation of the model was performed against several material test data from the literature with a simple one-element model. The model was also introduced into the human head FE model of THUMS v4.02 and validated against post-mortem human subject (PMHS) test data about brain displacements and intracranial pressures during head impacts.

A technique has been developed that uses unsteady flow simulation to evaluate mirror vibration quantitatively at the drawing stage. Studies made in actual driving tests of the contributions of different inputs to mirror vibration have confirmed that the contribution of fluid force is large, so a visualization of the structure of the external rearview mirror wake was done using PIV. The results made it clear that the vibration imparted to the mirror surface by air flow excites the natural vibration mode of the mirror surface, thereby causing the mirror to vibrate. Mirror vibration performance was evaluated by means of unsteady flow simulation using the moment PSD as a substitute characteristic. (The moment PSD was obtained by a frequency analysis of the changes over time in the moment generated in the mirror surface by the fluid force.) The results obtained through CFD show a high degree of correlation with those obtained in actual driving tests.

Several three-dimensional (3D) finite element (FE) models of the human body have been developed to elucidate injury mechanisms due to automotive crashes. However, these models are mainly focused on 50th percentile male. As a first step towards a better understanding of injury biomechanics in the small female, a 3D FE model of a 5th percentile female human chest (FEM-5F) has been developed and validated against experimental data obtained from two sets of frontal impact, one set of lateral impact, two sets of oblique impact and a series of ballistic impacts. Two previous FE models, a small female Total HUman Model for Safety (THUMS-AF05) occupant version 1.0ϐ (Kimpara et al., 2002) and the Wayne State University Human Thoracic Model (WSUHTM, Wang 1995 and Shah et al., 2001) were integrated and modified for this model development.

Flows around a forward facing step and a fence are simulated on structured grid to estimate aerodynamic noise by using direct simulation. Calculated results of sound pressure level show quantitatively good agreement with experimental results. To estimate aerodynamic noise from 3D complex geometry, a simplified side mirror model is also calculated. Averaged pressure distribution on the mirror surface as well as pressure fluctuations on the mirror surface and ground are simulated properly. However, calculated result of sound pressure level at a location is about 20dB higher than experiment due to insufficient spatial resolution. To capture the propagation of sound waves, more accuracy seems to be required.

In order to make more effective use of simulation, a new evaluation method on pressure loss has been developed. Using this method, the location and amount of pressure loss is clarified. As a result, enormous simulation data have been used effectively. The authors have confirmed the validity of the method about fundamental shapes.

In direct-injection (DI) gasoline engines, spray characteristics greatly affect engine combustion. For the rapid development of new gasoline direct-injectors, it is necessary to predict the spray characteristics accurately by numerical analysis based on the injector nozzle geometry. In this study, two-phase flow inside slit nozzle injectors is calculated using the volume of fluid method in a three-dimensional CFD. The calculation results are directly applied to the boundary conditions of spray calculations, of which the submodels are recently developed to predict spray formation process in direct injection gasoline engines. The calculation results are compared with the experiments. Good agreements are obtained for typical spray characteristics such as spray shape, penetration and Sauter mean diameter at both low and high ambient pressures. Two slit nozzle injectors of which the slit thickness is different are compared.

The purposes of this paper are to develop a new laboratory oxidation stability testing method and to clarify factors relative to the viscosity increase of engine oil. Polymerized products, obtained from the oil after a JASO M333-93 engine test, were found to consist mainly of carboxyl, nitrate and nitro compounds and to increase the oil viscosity. A good similarity between the JASO M333-93 test and the laboratory simulation test was found for the polymerized products. The products were obtained not by heating oil only in air but by heating oil while supplying a synthetic blowby gas consisting of fuel pyrolysis products, NO, SO2 and air. The laboratory test has also revealed that the viscosity increase depends on oil quality, organic Fe content and hydrocarbon composition in the fuel. Moreover, it has been found that blowby gas and organic Fe accelerate ZnDTP consumption and that aromatics concentration in the fuel correlates with the viscosity increase of oil.

The experiment-based droplet impinging breakup model was applied to a fan shaped spray and the impinging behavior was analyzed quantitatively. Evaluation of the quantitative results with validation tests verified the following. The model enables prediction of fan shaped spray thickness after impingement caused by the breakup of fuel droplets, which could not be represented with the Wall-Jet model, widely used at present. Fuel film movement on a wall is negligible when the injection pressure of the fan shaped spray is high and the spray travelling length is not too short. The proposed heat transfer coefficient between fuel film and the wall is too small to represent the vaporizing rate of the fuel film.

The drag reduction mechanism in newly developed low aerodynamic drag model car is investigated through numerical simulation. In order to deal with the computational domain around a three-dimensional complicated vehicle body, the method of overlaid grid system is employed. The results of computational case study on the body shape demonstrate that the lateral tapering near the rear end and the spats around the wheels bring better flow properties for drag reduction, such as the pressure recovery in the wake.

A one-dimensional (1-D) micro-kinetic reaction model with considering mass transport inside porous washcoat was developed to promote an effective development of multi-functional catalysts. The validation of this model has been done successfully through the comparison with a set of basic experiments. A numerical simulation study was conducted for the various catalyst configurations of three-way catalysts under Federal Test Procedure (FTP75) condition. It was found that a double layer type had a significant advantage in the total mass emissions, especially in NOx emissions. The reaction mechanisms in these catalysts were numerically clarified from the view point of detailed reaction dynamics. We concluded that the utilization of the numerical simulation with the detailed chemistry was effective for the optimization of catalyst design.

In this study, with the purpose of applying to the metal catalyst substrates, we have examined the feasibility of improving the light-off performance of a catalytic converter by enhancing heat transfer in the cells with heat-transfer promoters. Experimental and CFD analyses have been conducted to estimate heat transfer rates and pressure losses of the model cells with hemispherical protrusions. The analyses show that, by enhancing heat transfer of the cells, the cell density can be reduced keeping the catalytic performance in the steady state at the same level as that of conventional ones. As a result, the thermal mass of the substrate can be also reduced effectively without an increase of the pressure loss, and consequently the light-off performance of the catalytic converter can be improved noticeably.

The performances of some vehicle control systems are influenced by changes in the weight of the vehicle. In these systems, it is important to be able to estimate the weight without the need for special sensors. When we use physical models to do this, we have to provide estimates for two or more unknown parameters. In addition, since such a method is influenced by disturbances in the measured signals, it is difficult to maintain an acceptable level of accuracy. So, after analyzing the physical phenomena, we developed a new method that eliminates the influence of the disturbances from the measured signals and constructed an estimation system that has a minimum number of unknown parameters that was capable of providing a more accurate estimate of a vehicle weight. This method was applied to the braking force control of an automatic transmission and its efficacy was verified.

We propose a security-testing framework to analyze attack feasibilities for automotive control software by integrating model-based development with model checking techniques. Many studies have pointed out the vulnerabilities in the Controller Area Network (CAN) protocol, which is widely used in in-vehicle network systems. However, many security attacks on automobiles did not explicitly consider the transmission timing of CAN packets to realize vulnerabilities. Additionally, in terms of security testing for automobiles, most existing studies have only focused on the generation of the testing packets to realize vulnerabilities, but they did not consider the timing of invoking a security testing. Therefore, we focus on the transmit timing of CAN packets to realize vulnerabilities. In our experiments, we have demonstrated the classification of feasible attacks at the early development phase by integrating the model checking techniques into a virtualized environment.

This study investigates the reduction of the Blade Passing Frequency (BPF) noise radiated from an automotive engine cooling fans, especially in case of the fan with an eccentric shroud. In recent years, with the increase of HV and EV, noise reduction demand been increased. Therefore it is necessary to reduce engine cooling fan noise. In addition, as a vehicle trend, engine rooms have diminished due to expansion of passenger rooms. As a result, since the space for engine cooling fans need to be small. In this situation, shroud shapes have become complicated and non-axial symmetric (eccentric). Generally, the noise of fan with an eccentric shroud becomes worse especially for BPF noise. So it is necessary to reduce the fan BPF noise. The purposes of this paper is to find sound sources of the BPF noise by measuring sound intensity and to analyze the flow structure around the blade by Computational Fluid Dynamics (CFD).

This paper presents typical flow structures around a 60%-scale wind-tunnel model of a Formula One (F1) car, using planar particle image velocimetry (PIV). The customized PIV system is permanently installed in a wind tunnel to help aerodynamicists in the development loop. The PIV results enhance the understanding of the mean velocity field in the two-dimensional plane in some important areas of the car, such as the front-wheel wake and the underfloor flow. These real phenomena obtained in the wind tunnel also help maintain the accuracy of simulations using computational fluid dynamics (CFD) by allowing regular checking of the correlation with the real-world counterpart. This paper first surveys recent literature on unique flow structures around the rotating exposed wheel, mostly that on the isolated wheel, and then gives the background to F1 aerodynamics in the late 2000s.

Dynamic wind-tunnel tests of a simplified passenger car model were conducted using a two-degree-of-freedom model shaker. Time-resolved aerodynamic loads were derived from a built-in six-component balance and other sensors while the model underwent sinusoidal heaving and pitching motions at frequencies up to 8 Hz. The experimental results showed that frequency-dependent gains and phase differences between the model height/angle and the aerodynamic loads are in close agreement with those predicted by large-eddy simulation (LES) using an arbitrary Lagrangian-Eulerian (ALE) method. Based on these findings, transient aerodynamic loads associated with lateral motions were also estimated by LES analysis. Based on the above results, a full-unsteady aerodynamic load model was then derived in the form of a linear transfer function. The force and moment fluctuations associated with the vertical and lateral motions are well described by the full-unsteady aerodynamic load model.

This paper proposes a novel method of verifying comprehensive driver model used for the evaluation of driving safety systems, which is achieved by coupling the traffic simulation and the driving simulator (DS). The method consists of three-step procedure. In the first step, an actual driver operates a DS vehicle in the traffic flow controlled by the traffic simulation. Then in the next step, the actual driver is replaced by a driver model and the surrounding vehicle maneuvers are replayed using the recorded data from the first step. Then, the maneuver by the driver model is compared directly with the actual driver's maneuver along the simulation time steps.