Context: Since autonomous vehicles operate in an open context, their software components, including data-driven ones, have to reliably process inputs (e.g., obtained by cameras) in order to make safe decisions. A key challenge when providing reliable data-driven components is insufficient training data, which could lead to wrong interpretation of the environment, thereby causing accidents. Aim: The goal of our research is to extend available training data of data-driven components for safe autonomous vehicles using the example of traffic sign recognition. Method: We developed an approach to create realistic image augmentations of various quality deficits and applied them on the German traffic sign recognition benchmark dataset (GTSRB). Results: The approach results in images augmented with (any combination of) seven different quality deficits affecting traffic sign recognition (rain, dirt on lens, steam on lens, darkness, motion blur, dirt on sign, backlight) and considers dependencies between combined quality deficits and influences from other contextual information. Conclusion: Our approach can be used to obtain more comprehensive datasets, especially also including samples with quality deficits that are difficult to gather. By structuring the augmentation into a set of basic components, the approach can be adapted for other application domains (e.g., person detection).
Measurements of atmospheric parameters, such as wind velocity, air temperature and moisture, provide important information for a diverse set of fields, ranging from estimating energy output of wind farms to predicting extreme weather events to understanding urban climatology. Performing these measurements quickly, reliably and with high accuracy presents a yet unsolved challenge. Especially when working in or close to complex terrain, such as forests, hillsides or urban landscapes, no available system can properly perform such measurements. The Fraunhofer Institute for Physical Measurement Techniques IPM is developing a novel multispectral scanning LiDAR system. The goal is to simultaneously and accurately measure wind speed, air temperature and moisture over complex terrain for the first time. We present the current state of a scanner system for synchronized steering of multiple laser beams from different LiDAR units towards positions in the commonly visible intersection volume, subtending up to 7/8th of the full solid angle. We also present the state of an in-house developed Doppler Wind LiDAR and our current proposal for a combined wind, air temperature and water vapor LiDAR.
The discussion on whether and how to continue support for almost mature renewable electricity (RES-E) technologies, such as onshore wind and PV, has recently intensified. In this paper we analyze arguments in the literature in favor and against the phase-out of renewables support in the context of increasingly competitive RES-E technologies. We conclude that there are good reasons to continue dedicated RES-E policies beyond 2020 for those technologies. Dedicated RES-E support can provide a predictable, secure investment framework that lowers the risk premiums required by investors and therefore reduces the capital costs of RES-E. In addition, there are still significant cost reduction potentials for these technologies. The increased use of renewables has multiple socio-economic benefits in addition to climate change mitigation. These arguments are still valid when looking at the current market situation characterized by oversupply and low prices on both the CO2 market and some power markets in Europe. Since renewables are not the main reason for the current oversupply, it would not be effective to take actions towards restoring market equilibrium in the form of radical or overall phase-out of RES-E support.