研究目的
To present a decision-making architecture for automated driving that does not require detailed prior maps, using OpenStreetMap (OSM) for global route planning and automatically generated driving corridors adapted with a vision-based algorithm and grid-based approach for localization uncertainty.
研究成果
The proposed decision-making architecture successfully provides autonomous driving capabilities without detailed prior maps by integrating global route planning with OSM, vision-based road corridor adaptation, and grid-based localization uncertainty management. Future work will focus on interaction-aware motion prediction for dynamic objects in crowded urban environments.
研究不足
The architecture's reliance on OSM for global route planning may introduce inaccuracies due to the open nature of OSM data. The vision-based adaptation and grid-based uncertainty management require further validation in more complex and crowded urban environments.
1：Experimental Design and Method Selection:
The architecture integrates global, local, and HMI components for autonomous driving without detailed prior maps. It uses OSM for global route planning and automatically generates driving corridors adapted with a vision-based algorithm and grid-based approach for localization uncertainty.
2：Sample Selection and Data Sources:
The approach uses information from OpenStreetMap (OSM) and real-time data from a front camera and LiDAR for environment perception.
3：List of Experimental Equipment and Materials:
Instrumented vehicle with RTK DGPS receiver, stereo camera, four-layers LiDAR, onboard computer with Intel Core i7-3610QE and 8Gb RAM, Android-based tablet for HMI.
4：Experimental Procedures and Operational Workflow:
The system plans a global route using OSM, generates driving corridors, adapts them with a vision-based algorithm, and considers localization uncertainty with a grid-based approach. The local planner then uses these corridors to plan trajectories.
5：Data Analysis Methods:
The trajectory generation module evaluates path candidates within the road corridor, selecting the one that minimizes a given cost function, and computes a speed profile for comfort and safety.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
