MORPH-U: Robust Motion Planning for V2X-Enabled Autonomous Driving
A recent study presents MORPH-U, a motion planning system designed for autonomous vehicles that addresses uncertainties in V2X communication, including issues like delayed, lost, or manipulated messages, along with changes in dynamic maps. Developed using the CARLA simulator, this framework integrates data from LiDAR, radar, and cameras with V2X messages (CAM/DENM) to create a Local Dynamic Map (LDM). It initiates Hybrid-A* replanning when hazards or map alterations are confirmed to impact the planned trajectory. The approach employs a multi-objective strategy that optimizes tracking error, safety margin (minimum TTC), responsiveness, and smoothness, utilizing Pareto-frontier analysis for optimal point selection. To mitigate unsafe replanning due to erroneous V2X signals, MORPH-U features a Byzantine-inspired acceptance gate. The paper can be found on arXiv with the identifier 2605.07370.
Key facts
- MORPH-U is a motion planning framework for V2X-enabled autonomous driving.
- It addresses uncertainties like delayed, dropped, or forged V2X messages.
- Built on the CARLA simulator.
- Fuses LiDAR, radar, camera, and V2X data into a Local Dynamic Map (LDM).
- Uses Hybrid-A* replanning when validated hazards or map changes occur.
- Multi-objective formulation includes tracking error, safety margin, responsiveness, and smoothness.
- Pareto-frontier analysis selects optimal operating points.
- Includes a Byzantine-inspired acceptance gate to filter faulty V2X triggers.
- Paper available on arXiv with ID 2605.07370.
Entities
Institutions
- arXiv