The human voice is and remains the best form of communication. Therefore, spoken language interfaces to computers are a topic that has engaged engineers and speech scientists since the fifties. Once limited to the realm of science fiction, speech technology has now passed the threshold of practicality. The commercial deployment of these systems has already begun, although the applications are still specialized for certain purposes. This paper presents a speech recognition system that is particularly suited for drivers and passengers in a vehicle environment. The first chapter is describing a typical system design of today. Tens of thousands of these units are already in use. Thereafter a future system design will be described, in which the voice control of a navigation system is the most challenging problem. Finally the papers shows as an outlook, what comes next after the future. In particular, distributed speech processing systems will be used based on client-server architecture.
A finite element model of the human lower extremity has been developed in this study to simulate lower extremity behavior in frontal car crashes. Precise geometry of the human lower extremity and material properties of the hard and soft tissues were introduced to the model. The performance of the model was evaluated by comparing with dynamic loading test data using post mortem human subjects (PMHS). The comparison proved its ability to estimate dynamic responses of the human lower extremity. A study was conducted using the model to investigate possible factors of loading to the ankle and tibia. Force and moment were calculated with different time history profiles of footwell intrusion and pelvis motion. The results suggested that timing of maximum intrusion was important as well as its magnitude. It was also found that loading to the tibia could be affected not only by intrusion but also by pelvis motion.
This paper treats the study case of semi-trailer structures by using Finite Element Method (FEM). It has the main proposal to stablish an automatic procedure for model preparation, that means: meshing generation, design loads and constraints applications, for a family of semi-trailer from 15000 to 35000 dm3 capacity, automatically built by the computer program GERTAP with a very few user interference. For this reason, the program does not demand any FEM expertise so that engineers can focus main construction problems without excessive concerning about model theoretical characteristics and model mistakes. At present moment, we are able to develop static analysis, with use of equivalent accelerations, in order to compute weighting, braking and turning loads. Soon, in a very near future, we are going to apply dynamics analysis that will simulate the actual bad conditions of Brazilian roads, so that fatigue-cracking problems could be prevented in design stage.
Tyre structures are based on composite materials that constitute numerous layers, each providing specific properties to the tyre mechanic and dynamic behaviour. In principle, the understanding of the partial contributions of the individual layers requires knowledge of its mechanical properties. In case of non-availability of such critical information, it is difficult to perform tyre FE analyses. In the current work, a methodology is proposed to study the tyre static and dynamic behaviour to estimate its constituents properties based on the measured quasi-static responses of the tyre for certain specific loads. As a first step, a simplified tyre numerical model with standard rubber material properties is modeled that can substantively predict the necessary tyre static responses, i.e. radial, longitudinal and lateral stiffness. These responses are correlated with the physical tyre response that are measured using a kinematic and compliance (K&C) test rig in the laboratory.
The absence of combustion engine noise pushes increasingly attention to the sound generation from other, even much weaker, sources in the acoustic design of electric vehicles. The present work focusses on the numerical computation of flow induced noise, typically emerging in components of flow guiding devices in electro-mobile applications. The method of Large-Eddy Simulation (LES) represents a powerful technique for capturing most part of the turbulent fluctuating motion, which qualifies this approach as a highly reliable candidate for providing a sufficiently accurate level of description of the flow induced generation of sound.
With the electrification of road vehicles comes new demands on the cooling system. Not the least when it comes to noise. Less masking from the driveline and new features, as for example, cooling when charging the batteries drives the need for silent cooling fans. In this work a novel e-fan is studied in different generalized installations and operating conditions. The fans (a cluster configuration) are installed in a test rig where the operation could be controlled varying the speed, flow rate and pressure difference over the fan. On the vehicle side of the fan a generalized packaging space (similar to an engine bay for conventional vehicles) is placed. In this packaging space different obstruction can be placed to simulate the components and radiators used in the vehicle. Here generalized simple blocks in different configuration are used to provide well defined and distinct test cases.
In addition to the typical broadband noise character of wind noise, tonal noise phenomena can be much more disruptive, regardless of the overall interior noise quality of the vehicle. Whistling sounds usually occur by flow over sharp edges and resonant gaps, but can also be caused by the feedback of sound waves with laminar boundary layers or separation bubbles and the resulting frequency-selective growth of boundary layer instabilities. Such aeroacoustic feedback can e.g. occur at the side mirror of a vehicle and one compellingly needs the coupling of acoustic and flow field. A compressible large eddy simulation (LES) is in principle suitable but one has to take care of any numerical artifacts which can disturb the entire acoustic field. This paper describes the possibility to resolve aeroacoustic feedback with a commercial 2nd/3rd order finite volume CFD code.
Raising demands towards lightweight design paired with a loss of originally predominant engine noise pose significant challenges for NVH engineers in the automotive industry. From an aeroacoustic point of view, low frequency buffeting ranks among the most frequently encountered issues. The phenomenon typically arises due to structural transmission of aerodynamic wall pressure fluctuations and/or, as indicated in this work, through rear vent excitation. A possible workflow to simulate structure-excited buffeting contains a strongly coupled vibro-acoustic model for structure and interior cavity excited by a spatial pressure distribution obtained from a CFD simulation. In the case of rear vent buffeting no validated workflow has been published yet. While approaches have been made to simulate the problem for a real-car geometry such attempts suffer from tremendous computation costs, meshing effort and lack of flexibility.
Axial cooling fans are commonly used in electric vehicles to cool batteries with high heating load. One drawback of the cooling fans is the high aeroacoustic noise level resulting from the fan blades and the obstacles facing the airflow. To create a comfortable cabin environment in the vehicle, and to reduce exterior noise emission, a low-noise installation design of the axial fan is required. The purpose of the project is to develop an efficient computational aeroacoustics (CAA) simulation process to assist the cooling-fan installation design. This paper reports the current progress of the development, where the narrow-band components of the fan noise is focused on. Two methods are used to compute the noise source. In the first method the source is computed from the flow field obtained using the unsteady Reynolds-averaged Navier-Stokes equations (unsteady RANS, or URANS) model.
Integration of acoustic material concepts into vehicle design process is an important part of full vehicle design. The ability to assess the acoustic performance of a particular sound package component early in the design process allows designers to test various designs concepts before selecting a final products. This paper describes an innovative acoustic material concept which is easily integrated in a design process through the use of a database of Biot parameters. Biot parameters are widely used in the automotive industry to describe the physical interactions between the acoustics waves travelling through foams, fibers or metamaterials and the solid and fluid phase of these poro-elastic materials. This new acoustic material concept provides a combination of absorption, transmission loss and added damping on the panel it is attached to.
The properties of a polyurethane foam are greatly influenced by the addition of graphite particles during the manufacturing process, initially used as a fire retardant. These thin solid particles perturbate the nucleation process by generating bubbles in its immediate vicinity. The preponderance of work so far has focused on foams that are locally relatively homogeneous. We propose a model for locally inhomogeneous foams (including membrane effects) consisting of a random stack of spheres that permits one to represent certain pore size distribution functions. The cellular structure of the foam is obtained through a Laguerre tessellation and the solid skeleton determined from the minimization of surface energy (Surface Evolver). The structure of real foam samples is analyzed using X-ray computed tomography and scanning electron microscopy followed by image processing to create computerized three-dimensional models of the samples.