Optimize obstacle interactions at intersections and non-intersection roads for memory driving and urban NOA, including decision-making for deceleration and bypassing multiple obstacles.
Improve acceleration/deceleration and lane-changing strategies on highways and urban roads to enhance efficiency, success rates, and human-like behavior while minimizing unnecessary lane changes.
Ensure seamless integration of memory driving across urban roads, highways, and elevated roads.
Iterate on issues and improve generalization for mass production projects, achieving key performance metrics.
Job requirements
Experience in mass production of memory driving and urban NOA systems, or deep involvement in decision-making and planning development for related projects. Familiarity with decision-making and game-theoretic strategies (e.g., simulation-based reasoning, MCTS, Level-K algorithms).
Proficient in C++ and traditional data structure algorithms, with the ability to efficiently implement new features.
Strong communication skills, sense of responsibility, and teamwork spirit.
Understanding and insights into deep learning-based planning and prediction methods.