AVLocal: Safe AV Maneuvers on Local Roads

Description:

AVLocal is an innovative autonomous vehicle system designed specifically for rural and infrastructure-challenged environments. It combines real-time perception, map-independent localization, and tailored maneuver decision logic to navigate safely despite faded lane markings and lack of connectivity. Using multi-modal sensors (LiDAR, cameras, radar) and a hybrid control framework blending Model Predictive Control and Reinforcement Learning, AVLocal optimizes trajectories and implements fallback protocols to maintain safety under various environmental constraints.

Key Advantages:

• Operates effectively without HD maps by utilizing SLAM and GPS-IMU based localization.
• Adapts dynamically to adverse and variable environmental conditions with real-time perception.
• Custom maneuver logic tailored for rural road challenges ensures safer navigation.
• Hybrid control framework enhances robustness through combined predictive and learning-based approaches.
• Independent of network connectivity with optional V2X communication support for enhanced integration.
• Includes safety fallback protocols to manage operational limits securely.

Problems Solved:

• Enables AV navigation on rural roads with poor or faded lane markings.
• Overcomes lack of high-definition maps and unreliable network connectivity in underserved regions.
• Addresses environmental challenges such as variable weather and road conditions impacting sensor reliability.
• Reduces safety risks associated with autonomous driving in infrastructure-limited areas.

Market Applications:

• Autonomous vehicle manufacturers targeting rural and suburban markets.
• Rural transit agencies seeking safer and more reliable automated transport solutions.
• Logistics and delivery services operating in low-infrastructure regions.
• Research and development for next-generation autonomous navigation systems.