Autonomous Vehicular Systems: Architectural Strategies for Adaptive Multi-Objective Configuration

The dynamic reconfiguration of vehicular control and management systems to adapt to different scenarios, particularly those with conflicting design goals, remains a challenging task. In this context, we propose a reference architecture and a generic process to integrate advanced gain scheduling with a feedback loop that continually updates configuration lookup tables to generate scenario-related configurations at runtime. For demonstration purposes, our approach is applied to a Hyperloop vehicle's magnetic suspension system to guarantee simultaneously optimized accuracy of control, energy consumption and passenger comfort for a multitude of scenarios. Additionally, the feedback mechanism increases resilience against mechanical failures, marking an advancement in vehicular system adaptability and reliability.