arXiv subjects

Inga Jatzkowski

Publications and source records attributed to Inga Jatzkowski.

6 recordsLinked to original sources

arXiv2025

Synth It Like KITTI: Synthetic Data Generation for Object Detection in Driving Scenarios

An important factor in advancing autonomous driving systems is simulation. Yet, there is rather small progress for transferability between the virtual and real world. We revisit this problem for 3D object detection on LiDAR point clouds and propose a dataset generation pipeline based on the CARLA simulator. Utilizing domain randomization strategies and careful modeling, we are able to train an object detector on the synthetic data and demonstrate strong generalization capabilities to the KITTI dataset. Furthermore, we compare different virtual sensor variants to gather insights, which sensor attributes can be responsible for the prevalent domain gap. Finally, fine-tuning with a small portion of real data almost matches the baseline and with the full training set slightly surpasses it.

cs.CV↗

arXiv2022

Compensating for the Absence of a Required Accompanying Person: A Draft of a Functional System Architecture for an Automated Vehicle

A major challenge in the development of a fully automated vehicle is to enable a large variety of users to use the vehicle independently and safely. Particular demands arise from user groups who rely on human assistance when using conventional cars. For the independent use of a vehicle by such groups, the vehicle must compensate for the absence of an accompanying person, whose actions and decisions ensure the accompanied person's safety even in unknown situations. The resulting requirements cannot be fulfilled only by the geometric design of the vehicle and the nature of its control elements. Special user needs must be taken into account in the entire automation of the vehicle. In this paper, we describe requirements for compensating for the absence of an accompanying person and show how required functions can be located in a hierarchical functional system architecture of an automated vehicle. In addition, we outline the relevance of the vehicle's operational design domain in this context and present a use case for the described functionalities.

cs.HC↗

arXiv2021

A Taxonomy to Unify Fault Tolerance Regimes for Automotive Systems: Defining Fail-Operational, Fail-Degraded, and Fail-Safe

This paper presents a taxonomy that allows defining the fault tolerance regimes fail-operational, fail-degraded, and fail-safe in the context of automotive systems. Fault tolerance regimes such as these are widely used in recent publications related to automated driving, yet without definitions. This largely holds true for automotive safety standards, too. We show that fault tolerance regimes defined in scientific publications related to the automotive domain are partially ambiguous as well as taxonomically unrelated. The presented taxonomy is based on terminology stemming from ISO 26262 as well as from systems engineering. It uses four criteria to distinguish fault tolerance regimes. In addition to fail-operational, fail-degraded, and fail-safe, the core terminology consists of operational and fail-unsafe. These terms are supported by definitions of available performance, nominal performance, functionality, and a concise definition of the safe state. For verification, we show by means of two examples from the automotive domain that the taxonomy can be applied to hierarchical systems of different complexity.

eess.SY↗

arXiv2021

A Knowledge-based Approach for the Automatic Construction of Skill Graphs for Online Monitoring

Automated vehicles need to be aware of the capabilities they currently possess. Skill graphs are directed acylic graphs in which a vehicle's capabilities and the dependencies between these capabilities are modeled. The skills a vehicle requires depend on the behaviors the vehicle has to perform and the operational design domain (ODD) of the vehicle. Skill graphs were originally proposed for online monitoring of the current capabilities of an automated vehicle. They have also been shown to be useful during other parts of the development process, e.g. system design, system verification. Skill graph construction is an iterative, expert-based, manual process with little to no guidelines. This process is, thus, prone to errors and inconsistencies especially regarding the propagation of changes in the vehicle's intended ODD into the skill graphs. In order to circumnavigate this problem, we propose to formalize expert knowledge regarding skill graph construction into a knowledge base and automate the construction process. Thus, all changes in the vehicle's ODD are reflected in the skill graphs automatically leading to a reduction in inconsistencies and errors in the constructed skill graphs.

cs.AI↗

arXiv2020

Supporting Safe Decision Making Through Holistic System-Level Representations & Monitoring -- A Summary and Taxonomy of Self-Representation Concepts for Automated Vehicles

The market introduction of automated vehicles has motivated intense research efforts into the safety of automated vehicle systems. Unlike driver assistance systems, SAE Level 3+ systems are not only responsible for executing (parts of) the dynamic driving task (DDT), but also for monitoring the automation system's performance at all times. Key components to fulfill these surveillance tasks are system monitors which can assess the system's performance at runtime, e.g. to activate fallback modules in case of partial system failures. In order to implement reasonable monitoring strategies for an automated vehicle, holistic system-level approaches are required, which make use of sophisticated internal system models. In this paper we present definitions and an according taxonomy, subsuming such models as a vehicle's self-representation and highlight the terms' roles in a scene and situation representation. Holistic system-level monitoring does not only provide the possibility to use monitors for the activation of fallbacks. In this paper we argue, why holistic system-level monitoring is a crucial step towards higher levels of automation, and give an example how it also enables the system to react to performance loss at a tactical level by providing input for decision making.

eess.SY↗

arXiv2017

Towards a Skill- And Ability-Based Development Process for Self-Aware Automated Road Vehicles

The development of fully automated vehicles imposes new challenges in the development process and during the operation of such vehicles. As traditional design methods are not sufficient to account for the huge variety of scenarios which will be encountered by (fully) automated vehicles, approaches for designing safe systems must be extended in order to allow for an ISO~26262 compliant development process. During operation of vehicles implementing SAE Levels 3+ safe behavior must always be guaranteed, as the human driver is not or not immediately available as a fall-back. Thus, the vehicle must be aware of its current performance and remaining abilities at all times. In this paper we combine insights from two research projects for showing how a skill- and ability-based approach can provide a basis for the development phase and operation of self-aware automated road vehicles.

eess.SY↗