Michael Klar

Development of 77 GHz Radar Lens Antennas for Automotive Applications Based on Given Requirements

Lens antennas are well suited for automotive radar sensors. The development of these kinds of antennas in the field of automotive radar applications is presented. Based on the given requirements for the associated radar senior the development of the antenna is shown. Therefore a fast and accurate antenna simulation tool based on geometric optics (GO) and a field of view simulator has been developed. The simulation chain from simulating the antenna elements to the radar field of view is presented.

Development of a mid range automotive radar sensor for future driver assistance systems

Radar sensors are key components of modern driver assistance systems. The application of such systems in urban environments for safety applications is the primary goal of the project “Radar on Chip for Cars” (RoCC). Major outcomes of this project will be presented and discussed in this contribution. These outcomes include the specification of radar sensors for future driver assistance systems, radar concepts, and integration technologies for silicon-germanium (SiGe) MMICs, as well as the development and evaluation of a system demonstrator. A radar architecture utilizing planar antennas and highly integrated components will be proposed and discussed with respect to system specifications. The developed system demonstrator will be evaluated in terms of key parameters such as field of view, distance, and angular separability. Finally, as an outlook a new mid range radar (MRR) will be introduced incorporating several concepts and technologies developed in this project.

Estimating expected error rates of random forest classifiers: A comparison of cross-validation and bootstrap

Statistical learning has recently seen an expansion of applications in different areas of science, finance and industry, as it plays a great role within the fields of statistics, data mining and artificial intelligence. Hence, it intersects with areas of engineering and other disciplines as well. It is used for both regression and classification problems. Solving these problems usually involves building/training a model/classifier and validating its performance for a given task. In this paper we compare two resampling methods for assessment of a random forest classifier: k-fold cross-validation and bootstrap. We use these methods to estimate the generalization error and to create learning curves. Both methods yield similar results on our data. The most important requirement for good generalization error estimates of either method is that the used data sample (i.e. the training dataset) represents the unknown true distribution of the data. This requirement cannot always be met in practice and results of resampling methods have to be interpreted with care if it is violated.