MORPH-U: Multi-Objective Resilient Motion Planning for V2X-Enabled Autonomous Driving in High-Uncertainty Environments via Simulation

Researchers present MORPH-U, a simulation-based autonomous driving system that integrates Vehicle-to-Everything (V2X) communication with LiDAR/radar/camera sensors while implementing Byzantine-inspired safeguards against forged or delayed messages. The framework uses multi-objective optimization to balance safety, comfort, and responsiveness in high-uncertainty environments, demonstrating resilience against coordinated false-message attacks.

MORPH-U addresses a critical vulnerability in next-generation autonomous vehicle architectures: the security and reliability gap introduced by V2X communication integration. While V2X enables vehicles to share hazard warnings and map updates beyond line-of-sight, it creates new attack surfaces through message delays, dropouts, and forgery. The researchers tackle this by developing a closed-loop system that validates V2X inputs against onboard sensor data before triggering replanning decisions.

The technical contribution centers on three innovations: a Local Dynamic Map that fuses multi-modal sensor inputs with V2X data streams, a Hybrid-A* replanning algorithm responsive to map changes within real-time constraints, and a lightweight Byzantine-inspired acceptance gate that uses quorum voting and sensor-based veto mechanisms. The multi-objective optimization formulation reveals explicit trade-offs between tracking accuracy, safety margins (measured by time-to-collision), responsiveness to new hazards, and ride smoothness.

For the autonomous vehicle industry, this research validates that V2X integration requires dual-layer validation rather than naive sensor fusion. The experimental results demonstrating robustness under complete false-DENM injection (100% attack saturation) provide evidence that Byzantine-resistant architectures can maintain safety even under adversarial conditions. This has implications for regulatory frameworks and insurance models, which currently lack clarity on V2X trust assumptions.

The CARLA simulation-based approach enables reproducible benchmarking but leaves open questions about real-world performance with actual communication latencies and sensor noise distributions. Future work should focus on field testing with production V2X hardware and exploring how Pareto-frontier tuning performs across diverse driving scenarios.

• →MORPH-U integrates V2X communication with sensor fusion while defending against forged or delayed messages using Byzantine-inspired acceptance gates
• →Multi-objective Pareto optimization reveals controllable trade-offs between safety margins, tracking accuracy, responsiveness, and passenger comfort
• →System maintains safety under 100% false-message injection attacks through quorum-based validation and onboard sensor veto mechanisms
• →Hybrid-A* replanning responds to map changes within real-time budgets, enabling dynamic route adaptation during trips
• →CARLA-based simulation provides reproducible benchmarking for V2X-augmented autonomous driving, though field validation remains pending