Hochschule Konstanz University of Applied Sciences
Hochschulschriftenserver der HTWG KonstanzNot a member yet
2319 research outputs found
Sort by
UHI risk mapping and assessment at street level using mobile mapping data and AI-based data analysis
No full textInfolge des Klimawandels werden Hitzeperioden häufiger und intensiver, was insbesondere in Städten zu einer Überwärmung des Straßenraums führt. Erhöhte Gesundheitsrisiken für vulnerable Gruppen sowie eine Minderung der Aufenthalts- und Lebensqualität sind die Folgen. Für die Stadtplanung ergibt sich die Notwendigkeit, dem Urban-Heat-Island-(UHI-)Effekt durch geeignete Klimaanpassungsmaßnahmen zu begegnen. Bisherigen Ansätzen zur Lokalisierung überwärmungsgefährdeter Bereiche fehlt oft die Detailtiefe, um einen direkten Straßenbezug herzustellen, Ursachen zu analysieren und geeignete Anpassungsmaßnahmen im Straßenraum abzuleiten. In diesem Beitrag wird daher ein Ansatz vorgestellt, der die Daten eines Mobile-Mapping-Systems nutzt, um UHI-Risikobereiche im städtischen Straßennetz präzise zu kartieren und zu bewerten. Das Bewertungskonzept ist so ausgelegt, dass gezielt Maßnahmen zur Verbesserung des Mikroklimas empfohlen werden können
Solgenia - A test vessel toward energy-efficient autonomous water taxi applications
No full textAutonomous surface vessels are a promising building block of the future’s transport sector and are investigated by research groups worldwide. This paper presents a comprehensive and systematic overview of the autonomous research vessel Solgenia including the latest investigations and recently presented methods that contributed to the fields of autonomous systems, applied numerical optimization, nonlinear model predictive control, multi-extended-object-tracking, computer vision, and collision avoidance. These are considered to be the main components of autonomous water taxi applications. Autonomous water taxis have the potential to transform the traffic in cities close to the water into a more efficient, sustainable, and flexible future state. Regarding this transformation, the test platform Solgenia offers an opportunity to gain new insights by investigating novel methods in real-world experiments. An established test platform will strongly reduce the effort required for real-world experiments in the future
Autonomous Driving Control Using Parallel Deep Reinforcement Learning Algorithms in Continuous Spaces
No full textAs one of many fields within today’s Artificial Intelligence, Deep Reinforcement Learning (DRL) has demonstrated significant advancements in solving complex decision-making tasks. As opposed to supervised and unsupervised learning, a DRL model learns by exploring an environment, collecting experiences, and maximizing a reward gained by performing specific actions in specific states. In this thesis, the development of various RL algorithms is discussed, with a primary focus on the Deep Deterministic Policy Gradient (DDPG) algorithm. Among many reinforcement learning algorithms, DDPG stands out as a powerful approach for tackling continuous action spaces such as those found in robotics and autonomous driving control. However, the computational demands of these algorithms present a significant challenge, particularly when applied to environments with high-dimensional state and action spaces. To address this issue, this thesis proposes a parallelized version of the DDPG algorithm
Faster-than-real-time Simulation of Multi-group Pedestrian Flow
No full textSimulations of pedestrian flow are widely used for design guidelines during the planning of buildings and large events. In this article we discuss how faster-than-real-time simulations of pedestrian flow can contribute to dynamic control strategies of the crowd in order to ensure safety live during an event. For this purpose we present a continuous model for multi-group pedestrian flow. The model consists of the continuity equation for the density of pedestrians and a Helmholtz type equation providing the fastest path to the exit. It is derived by regularizing Hughes’ macroscopic model for pedestrian flow and is solved using the finite element method, which enables simulations at short runtimes. The model allows the definition of multiple groups with different fundamental diagrams or destination preferences. We show that this model is able to reproduce qualitatively the expected behavior of crossing pedestrian crowds. Its main advantage is its simulation speed, making it ideal for real-time crowd management at large events
Arbeitsrecht Fälle und Schemata für Dummies
No full textDas in der Zweitauflage erschienene grundlegend überarbeitete und ergänzte Werk enthält die wichtigsten arbeitsrechtlichen Prüfungsschemata sowie 20 Übungsfälle mit exemplarischen Lösungen zur Selbstkontrolle
Forschung und Transfer Jahresbericht 2024
No full textÜbersicht über Drittmitteleinnahmen im Bereich Forschung und Transfer sowie wissenschaftliche Publikationen inkl. abgeschlossener Promotionen
