Karl-Heinz Siedersberger

Situation aspect modelling and classification using the Scenario Based Random Forest algorithm for convoy merging situations

Advanced Driver Assistance Systems (ADAS) require an understanding of complex traffic situations. Such complexity can be handled by decomposing traffic situations into analyzeable subsets called Situation Aspects (SA). Since lots of situation analyzing problems result in classification tasks, the Scenario Based Random Forest (SBRF) algorithm is introduced into the field of ADAS situation analysis research. This classification method is designed to handle feature sets that develop over time and classification results that can be judged only by using complete scenarios instead of single time snap-shots. Furthermore, it has the advantage of using the out of bag (oob) estimation technique in order to perform feature selection. The problem of detecting a convoy merging traffic situation in real traffic scenarios serves as example to show the process of situation aspect modelling, feature selection and classification using the above mentioned methodology. It is demonstrated how the challenge of labelling changing SA can be solved using an undefined transition class and how this effects classification results. Because unbalanced data sets often occur in ADAS situation analysis, results on over- and downsampling strategies are described as well.

Strategy and architecture of a safety concept for fully automatic and autonomous driving assistance systems

As drivers back out of the driving task, when transported automatically by an intelligent car for a longer time, they are not always able to react properly, if a driver take over request occurs. This paper presents two ways, how to deal with this problem within the scope of a functional safety concept. Thereto, the difference between fully automatic and autonomous driving assistance systems is explained. Afterwards two different strategies to reach a safe state in consequence of a system boundary crossing are proposed. In the first case the fall back state is reached by a driver take over, in the second case by an automatic, active fail-safe mechanism. Subsequently the necessary components for monitoring and reaching a safe state and their embedment in a basic, functional architecture of a driving assistance system are described. In this context, special regard is paid to aspects of redundancy as well. In the end it is concluded, that the safety concept proposed here is crucial for guaranteeing enduring safety in an automatically driving car and in consequence for making automatic driving functions commercially ready for serial production.

Automatic Emergency Steering with Distracted Drivers: Effects of Intervention Design

Driver inattention is reported to be one of the most prominent contributing factors to crashes. Modern vehicles feature sensor equipment able to detect an imminent collision, potentially permitting advanced driver assistance systems (ADAS) to cope for such human error. Steering interventions, however, make high demands on the human-machine-interaction. Unlike in autonomous emergency braking, conflicting driver input cannot be omitted. Three different ADAS configurations for an automatic emergency steering intervention with small lateral offset were tested against an unassisted baseline condition in a driving experiment with distracted drivers. The results suggest an influence of feedback modalities and actuator choice.

Timestamping and latency analysis for multi-sensor perception systems

This paper describes methods and tools for the analysis of measurement timestamps and time delays of different environmental perception sensors for Advanced Driver Assistance Systems (ADAS) in order to allow a proper sensor data interpretation. These sensors are classified according to the way they timestamp their measurements. Thereby, three classes are defined: sensors with global timestamps, sensors with relative timestamps and sensors without measurement timestamps. This paper describes the concept and tools for analyzing and setting the timestamp of sensors from class I and II. For the classI, a global UTC1-master time generator allows the creation of a relationship between the measurement and the acquisition timestamps. For the class II, a synchronization signal is periodically sent to the sensors which timestamp their data relative to this reference signal. The used tools for this analysis were thoroughly analyzed and verified.

Validation of driving behavior in the Vehicle in the Loop: Steering responses in critical situations

Combining real driving experience with the safety and replicability of a simulation, the Vehicle in the Loop (VIL) setup potentially qualifies as an ideal tool for the development and evaluation of safety-related driver assistance systems. Previous studies have assessed this high-fidelity simulators validity for longitudinal driving behavior. Aiming to validate the VIL setup with regard to lateral driving behavior, an experiment was conducted comparing steering behavior of 48 participants throughout the same critical traffic scenario experienced in a real driving environment and in the VIL simulation. Steering behavior was found to be less intense in the virtual driving environment. It was, however, affected equally in both driving environments by driver-independent steering wheel torque, demonstrating relative validity of the VIL setup with regard to steering responses in challenging situations.

Stereo and photometric image sequence interpretation for detecting negative obstacles using active gaze control and performing an autonomous jink

Driving cross-country, the detection and state estimation relative to negative obstacles like ditches and creeks is mandatory for safe operation. Very often, ditches can be detected both by different photometric properties (soil vs. vegetation) and by range (disparity) discontinuities. Therefore, algorithms should make use of both the photometric and geometric properties to reliably detect obstacles. This has been achieved in UBMs EMS-Vision System (Expectation-based, Multifocal, Saccadic) for autonomous vehicles. The perception system uses Sarnoffs image processing hardware for real-time stereo vision. This sensor provides both gray value and disparity information for each pixel at high resolution and framerates. In order to perform an autonomous jink, the boundaries of an obstacle have to be measured accurately for calculating a safe driving trajectory. Especially, ditches are often very extended, so due to the restricted field of vision of the cameras, active gaze control is necessary to explore the boundaries of an obstacle. For successful measurements of image features the system has to satisfy conditions defined by the perception expert. It has to deal with the time constraints of the active camera platform while performing saccades and to keep the geometric conditions defined by the locomotion expert for performing a jink. Therefore, the experts have to cooperate. This cooperation is controlled by a central decision unit (CD), which has knowledge about the mission and the capabilities available in the system and of their limitations. The central decision unit reacts dependent on the result of situation assessment by starting, parameterizing or stopping actions (instances of capabilities). The approach has been tested with the 5-ton van VaMoRs. Experimental results will be shown for driving in a typical off-road scenario.

Anforderungen an ein Referenzsystem für die Fahrzeugortung

Im Hinblick auf einen automatisierten Verkehr werden Fahrerassistenzsysteme in Zukunft immer starker in die Fahrzeugfuhrung eingreifen. Derartige Systeme stutzen sich vor allem auf die Umfeldwahrnehmung und insbesondere auf die Kenntnis der eigenen Fahrzeugposition. Hierbei entsteht der Bedarf, die spezifizierte Messqualitat der Ortungssysteme zu verifizieren. In einem gemeinsamen Forschungsprojekt der TU Braunschweig und Audi wurden strukturelle und parametrische Anforderungen an ein Referenzsystem kategorisiert, aus dem sich fur den jeweils betrachteten Anwendungsfall konkrete Anforderungen ableiten lassen.

Determination of Complex Permittivity of LRR Radome Materials Using a Scalar Quasi-Optical Measurement System

We developed a low-cost quasi-optical measurement system to determine the complex permittivity in the E-band (from 60 to 90 GHz). The evaluation is done in a non-destructive way and can be used for all kinds of single-layered and multi-layered dielectric materials. The method is based on measurements of the scalar transmission coefficient through planar samples for several angles of incidence and two different polarization states (parallel and perpendicular to the plane of incidence). A numerical optimization technique is used to derive the complex permittivity from the measured coefficients. The method utilizes a physical model of a dielectric slab of known thickness, which assumes that a plane wave is incident on the surface of the dielectric material. Measurement results are presented, which are in good agreement with data from the literature.