Efficient Hybrid Hard--Soft Signal Temporal Logic MPC for Driving under Multimodal Uncertainty

This research proposes a hybrid hard-soft Signal Temporal Logic (STL) MPC framework for motion planning in automated vehicles, ensuring safety, compliance with traffic rules, and comfort under multimodal uncertainty. The method uses explicit hard constraints for critical safety requirements and soft STL terms to optimize clearance, comfort, and interaction quality. Closed-loop evaluations show high success rates with reduced planning latency compared to unoptimized formulations.