引导式测验 · 中等 等级

Wayve 和端到端驾驶模型测验

比较 Wayve 的端到端驾驶模型、无地图 AV2.0 方法、可控 GAIA-2 世界模型和与车辆无关的平台。

问题 1 的 8

Wayve 的端到端驾驶模型将什么映射为驾驶输出?

预览每个问题8 问题

当您参加测验时,就会出现答案和解释。

  1. Wayve 的端到端驾驶模型将什么映射为驾驶输出?
    • Raw sensor inputs
    • A prebuilt high-definition map only
    • A human operator's verbal instructions only
    • A fixed route table without vehicle observations
  2. Wayve 的 AV2.0 方法强调了哪种映射策略?
    • A hand-built HD map for every street
    • Mapless autonomy that does not rely on high-definition maps
    • LiDAR-only navigation with no learned model
    • A fixed route table programmed for each trip
  3. Wayve 堆栈中的 GAIA-2 是什么?
    • LiDAR 传感器模型
    • 物理机器人出租车平台
    • 模拟驾驶视频的生成世界模型
    • 监管认证标准
  4. Wayve 主要计划如何将其技术商业化?
    • 向供应商销售激光雷达装置
    • 独家运营自己的全市机器人出租车车队
    • 构建消费者地图应用程序
    • 向汽车制造商授权其驾驶软件
  5. 单一端到端驾驶网络的主要缺点是什么?
    • 比模块化管道更难解释和调试
    • 它不能使用相机
    • 它不需要训练数据
    • 它仅适用于激光雷达
  6. What learning approach does Wayve describe for its AV2.0 system?
    • Training only from hand-coded driving rules
    • Self-supervised learning from raw, unlabeled driving data
    • A fixed route table for every city
    • A speech-recognition model trained on transcripts
  7. What driving factors can GAIA-2 control when generating a scenario?
    • Only the color of the vehicle logo
    • Only the final steering angle
    • Ego-vehicle actions, dynamic agents, and scene features
    • The legal approval status of a car
  8. What does Wayve mean by a sensor-agnostic AI Driver platform?
    • Every vehicle must use the same camera hardware
    • The system can drive without any sensor input
    • It requires an HD map for every road
    • Automakers can integrate it with sensor and compute choices suited to their vehicles