go back

Hybrid Probabilistic Logic Programming for Mission Design in Multimodal Mobility

Simon Kohaut, "Hybrid Probabilistic Logic Programming for Mission Design in Multimodal Mobility", Technical University of Darmstadt, 2022.


Reasoning on subjective observations of the environment to navigate through complex and dynamic scenarios is a fundamental concept to human behavior. Hence, over the course of history, a vast landscape of approaches to formalize and automize inference has emerged. From propositional to higher-order logic and from simple stochastic measures to robust statistics, the power of both symbolic and numeric methods for reasoning in discrete and continuous worlds has drastically grown. Hereby, our perspective on embedding fully autonomous systems in complex, human inhabited environments has improved steadily over time. To make this a reality, automatic reasoning, interfacing with human desires and imposing strict laws on an agent's behavior, is key to allow for safe and secure deployment. Especially in modern mobility applications, a multitude of systems needs to utilize limited resources to move goods and people without violating traffic rules or safety measures. This work tackles the question on how we can intertwine novel traffic systems for autonomous agents with recent advances in statistical relational AI to obtain a transparent system for mission design that is able to adapt to changing rules and user preferences.

Download Bibtex file Per Mail Request