Hesai Vice President Yang Xiansheng presented an innovative robot visual perception architecture at ICRA 2026. This architecture achieves natural alignment of depth detection and RGB information at the physical level, outputting without post-calibration, thereby significantly improving the quality of robot training data and real-time perception capabilities. Yang Xiansheng pointed out that the current bottleneck in robot perception lies in the inability to achieve millimeter-level depth perception. Hesai's SPAD technology route, through high integration, achieves a generational advantage in performance, breaking the impossible triangle between stability, distance, and accuracy, and accelerating the evolution of physical AI. AI
影响 This architecture is expected to solve the robot perception bottleneck, improve data quality and training efficiency, thereby accelerating the evolution of embodied intelligence and physical AI.
排序理由 文章详细介绍了在会议上展示的机器人视觉和感知新架构，重点关注其技术优势和对人工智能发展的潜在影响。
AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →
