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。
- 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
- 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
- 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
- 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
- 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
- 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
- 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
- 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