Robust and Efficient Vehicle Automation
Robust software modules that continuously evaluate their quality form the basis for safely operating vehicles. This is especially true in challenging situations or in the event of a fault.
In autotech.agil, we are researching the development of robust and modular vehicle automation systems. Modular architectures enable software and hardware components to be updated and expanded flexibly. The goal is to develop an open architecture that can be used in different types of vehicles and mobility concepts, while ensuring a high level of safety, energy efficiency and resilience.
A particular focus is placed on modular and service-oriented software architectures that enable the exchange and updating of functions even during runtime. This is an important prerequisite for learning systems that benefit from continuous improvement. Standardization processes are sought to facilitate compatibility and integration into existing systems.
A central aspect in autotech.agil is the establishment of a “quality vector” that describes the capabilities and states of individual modules and thus enables safe behavior in the event of a fault. In addition, modules can be dynamically orchestrated or modularly secured using the quality vector. This not only increases the safety and reliability of the systems, but also their adaptability to different conditions.
Besides developing and integrating this quality vector along the entire processing chain of the automation system, we are also working in autotech.agil on improved and more reliable methods for localization and environment perception, as well as the consideration of probabilistic information in the motion planning of the vehicle. This also includes the investigation of novel sensor technologies based on corresponding AI data processing methods, for example with neuromorphic hardware. Based on a sensor module for the near field, research is also being conducted into automation at low speeds, for example to implement valet parking in vehicles with high-quality assistance sensors.
Furthermore, the interconnection of vehicles with the cloud is necessary to model a live digital twin of traffic and to enable cooperative functions. This increases traffic safety and enables new mobility concepts such as “collectively assisted” vehicles.
Overall, the aim of robust and modular vehicle automation is to establish a flexible, efficient and, in particular, safe architecture and to sustainably transform the automotive industry by introducing new methods and concepts.
