Search Results

Highly automated driving (HAD) is under rapid development and will be available for customers within the next years. However the evidence that HAD is at least as safe as human driving has still not been produced. The challenge is to drive hundreds of millions of test kilometers without incidents to show that statistically HAD is significantly safer. One approach is to let a HAD function run in parallel with human drivers in customer cars to utilize a fraction of the billions of kilometers driven every year. To guarantee safety, the function under test (FUT) has access to sensors but its output is not executed, which results in an open loop problem. To overcome this shortcoming, the proposed method consists of four steps to close the loop for the FUT. First, sensor data from real driving scenarios is fused in a world model and enhanced by incorporating future time steps into original measurements.

In the paper some methods are presented, with the corresponding practical examples, related to MonteCarlo (MC) techniques for thermal model/test correlation purposes. The MonteCarlo techniques applied to model correlation are intended to be used as an alternative to empirical ‘manual’ correlation techniques, gradients methods, matrix methods based on least square fit minimization. First of all, Design Of Experiments (DoE) tools are used to determine the model response to uncertain parameters and the confidence level of such a response. A sensitivity map is built, allowing the design of the test to maximize the response of the system to the uncertain parameters. Techniques derived from the extreme statistics are used to extrapolate data beyond test limits, with a sufficient confidence in the queue behaviour.

Transport activities contribute significantly to the air pollution and its impact on emissions is a key element in the evaluation of any transport policy or plan. Calculation of emissions has therefore gained institutional importance in the European Community. To obtain emission factors several methods make use of only vehicle mean velocity, which can be easily obtained by vehicle flow and density in the road. Recently in ARTEMIS project by Rapone et al. (2005–2007) a meso scale emission model, named KEM (Kinematic Emission Model), able to calculate emission factor has been developed. This model is based on a new statistical methodology, capable to consider more attributes than the simple mean speed to characterize driving behaviour. An interesting approach to determine the exact mix of driving cycles is represented by the use of microscopic traffic simulation models that could be used to avoid the very expensive costs of experimental campaigns needed to obtain real driving cycle.

Transport activities contribute significantly to air pollution. For this reason any policy or plan, carried out by administration or institution, requires the assessment of its impact on the emissions. To assess the overall pollutant production from transport, it is necessary to calculate emission factors. For this aim several methods exist which only use the average speed of the traffic stream, which can be theoretically obtained by vehicles flow and density on the road. Recently, a new statistical approach has been developed capable to consider more attributes than the simple mean speed to characterize driving behaviour, not only in the determination of driving cycles but also in the emission modelling. In this context, a meso scale emission model, named KEM, Kinematic Emission Model, able to calculate emission factor was developed. However, it is necessary to consider that the input to this model is, in any case, the driving cycle.

Two models for the forecast of road traffic emissions, independently developed in parallel, are comparatively presented and assessed: EPROG developed by BMW and enlarged by VDA for a national application (Germany) and FOREMOVE, developed for application on European Community scale. The analysis of the methodological character of the two algorithms proves that the models are fundamentally similar with regard to the basic calculation schemes used for the emissions. The same holds true as far as the significant dependencies of the emission factors, and the recognition and incorporation of the fundamental framework referring to traffic important parameters (speeds, mileage and mileage distribution etc) are concerned.

The increasing number of local control units in automotive systems led to growing emphasis on developing and using multiplex systems. For reasons of price and robustness the use of asynchronous and slow multiplex systems is preferred. Since the communication volume now reaches critical dimensions in peak load situations during the use of those systems, new concepts on different communication levels have to be developed. Due to the use of many different message types (wide range of message length) and the statistical dependence of the communication behaviour of control units (e.g. question-answer-combinations), the application of standard methodologies is only partly suitable for a performance analysis of automotive multiplex systems.

The increasing complexity of communication protocols for asynchronous multiplex systems requires the use of simulation during the optimisation of these protocols or the integration of other control units. Consideration of realistic communication behaviour of the connected control units is essential for performance analysis of multiplex systems. For a first pass, the use of simple statistical distributions (e.g. Poisson distribution) is suitable to get some simulation results. A better way to get realistic results is the approximation of empirical communication data through the use of more complex statistical distribution (e.g. mixed Erlang distributions). In this paper several approaches for the approximation of empirical data are presented. Beside simple statistical distributions (with one parameter), the use of more complex statistical distributions is discussed and methods for the identification of their parameters are presented.

We present a framework for the robust optimization of the heat flux distribution for an anti-ice electro-thermal ice protection system (AI-ETIPS) and iced airfoil performance analysis under uncertain conditions. The considered uncertainty regards a lack of knowledge concerning the characteristics of the cloud i.e. the liquid water content and the median volume diameter of water droplets, and the accuracy of measuring devices i.e., the static temperature probe, uncertain parameters are modeled as uniform random variables. A forward uncertainty propagation analysis is carried out using a Monte Carlo approach. The optimization framework relies on a gradient-free algorithm (Mesh Adaptive Direct Search) and three different problem formulations are considered in this work. Two bi-objective deterministic optimizations aim to minimize power consumption and either minimize ice formations or the iced airfoil drag coefficient.

The German funded project ARAMiS included work on several demonstrators one of which was a multicore approach on large scale software integration (LSSI) for the automotive domain. Here BMW and Audi intentionally implemented two different integration platforms to gain both experience and real life data on a Hypervisor based concept on one side as well as using only native AUTOSAR-based methods on the other side for later comparison. The idea was to obtain figures on the added overhead both for multicore as well as safety, based on practical work and close-to-production implementations. During implementation and evaluation on one hand there were a lot of valuable lessons learned about multicore in conjunction with safety. On the other hand valuable information was gathered to make it finally possible to set up a cost model for estimation of potential overhead generated by different integration approaches for safety related software functions.

The dynamics and evolution of turbulent structures inside an engine-like geometry are investigated by means of Large Eddy Simulation. A simplified configuration consisting of a flat-top cylinder head with a fixed, axis-centered valve and low-speed piston has been simulated by the finite volume CFD code OpenFOAM®; the standard version of the software has been extended to include the compressible WALE subgrid-scale model, models for the generation of synthetic turbulence, some improvements to the mesh motion strategy and algorithms for LES data post-processing. In order to study both the initial transient and the quasi- steady operating conditions, ten complete engine cycles have been simulated. Phase and spatial averages have been performed over cycles three to ten in order to extract first and second moment of velocity; these quantities have then been used to validate the numerical procedure by comparison against experimental data.

Metallic open-cell foams have proven to be valuable for many engineering applications. Their success is mainly related to mechanical strength, low density, high specific surface, good thermal exchange, low flow resistance and sound absorption properties. The present work aims to investigate three principal aspects of real foams: the geometrical characterization, the flow regime characterization, the effects of the pore size and the porosity on the pressure drop. The first aspect is very important, since the geometrical properties depend on other parameters, such as porosity, cell/pore size and specific surface. A statistical evaluation of the cell size of a foam sample is necessary to define both its geometrical characteristics and the flow pattern at a given input velocity. To this purpose, a procedure which statistically computes the number of cells and pores with a given size has been implemented in order to obtain the diameter distribution.