近日，前Waymo首席执行官约翰·克拉夫西克（John Krafcik）公开批评特斯拉依赖纯视觉方案（即仅使用摄像头）实现自动驾驶的做法。他指出，这种策略在技术上存在明显局限，尤其是在复杂或低能见度的驾驶环境中，缺乏激光雷达（LiDAR）等传感器的支持，难以确保系统安全可靠。克拉夫西克强调，真正的自动驾驶必须建立在多传感器融合的基础上，包括摄像头、雷达和激光雷达，以提供冗余和更高的环境感知精度。他还质疑特斯拉将“高级驾驶辅助系统”（如Autopilot和FSD）宣传为接近完全自动驾驶的做法，认为这可能误导消费者并带来安全隐患。作为谷歌旗下自动驾驶公司Waymo的前掌舵人，克拉夫西克一直主张采用更保守但更安全的技术路径。他的言论再次引发了业界对自动驾驶技术路线之争的关注：是追求低成本、可快速部署的纯视觉方案，还是坚持高成本但更可靠的多传感器融合方案？目前，大多数主流自动驾驶公司（如Cruise、Argo AI、百度Apollo等）仍选择后者，而特斯拉则坚定走纯视觉路线，试图通过AI算法弥补硬件上的“缺失”。这场技术理念的分歧，或将深刻影响未来自动驾驶的发展方向。
Recently, former Waymo CEO John Krafcik publicly criticized Tesla’s reliance on a vision-only approach—using only cameras—for its autonomous driving system. He argued that this strategy has clear technical limitations, especially in complex or low-visibility driving conditions, where the absence of sensors like LiDAR makes it difficult to ensure safety and reliability. Krafcik emphasized that true autonomy must be built on multi-sensor fusion, combining cameras, radar, and LiDAR to provide redundancy and higher environmental perception accuracy. He also questioned Tesla’s marketing of its “advanced driver-assistance systems” (such as Autopilot and Full Self-Driving) as being close to full autonomy, warning that this could mislead consumers and pose safety risks. As the former head of Waymo, Google’s self-driving unit, Krafcik has long advocated for a more cautious yet safer technological path. His remarks have reignited industry debate over autonomous driving strategies: should companies pursue a low-cost, rapidly deployable vision-only system, or stick with a more expensive but reliable multi-sensor approach? Currently, most major autonomous driving firms—including Cruise, Argo AI, and Baidu Apollo—favor the latter, while Tesla remains committed to its vision-only route, betting that advanced AI algorithms can compensate for the lack of additional hardware. This fundamental philosophical divide may significantly shape the future trajectory of autonomous vehicle development.
原创文章，作者：admin，如若转载，请注明出处：https://ktez.cn/18748.html