Scholarlink

AI-based perception systems require large amounts of data for training and validation, particularly data from rare or even hazardous real-world scenarios, to improve the reliability and safety of AI perception systems in actual driving conditions. These scenarios are known as key scenarios or corner cases. Collecting key scenarios through real-world testing is often economically ineffcient, risky, and yields limited data.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Real-Virtual Scene Augmentation and Reinforcement Learning Exploration

  • 章节

  • 发表日期：

    2026年

查看全文
  • Kun Gao

摘要

AI-based perception systems require large amounts of data for training and validation, particularly data from rare or even hazardous real-world scenarios, to improve the reliability and safety of AI perception systems in actual driving conditions. These scenarios are known as key scenarios or corner cases. Collecting key scenarios through real-world testing is often economically ineffcient, risky, and yields limited data.