自动驾驶的激光雷达和纯视觉方案各有什么优劣，为什么很多车企仍在争论“要不要激光雷达”？
What are the pros and cons of LiDAR versus pure vision for autonomous driving, and why do automakers still debate whether to use LiDAR?
激光雷达直接测距、夜间和逆光都稳，但成本高、怕恶劣天气；纯视觉便宜且像人眼，却依赖算法和算力去“脑补”深度。成本与冗余安全的权衡让两派长期对立。
LiDAR measures distance directly and works in darkness and backlight but is costly and weather-sensitive; vision is cheap and human-like yet relies on algorithms and compute to infer depth. The trade-off between cost and redundancy keeps the camps opposed.
难度 专业 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Maurer、Gerdes、Lenz & Winner，《自动驾驶：技术、法律与社会视角》（Springer, 2016）
Maurer, Gerdes, Lenz & Winner, 'Autonomous Driving: Technical, Legal and Social Aspects' (Springer, 2016)
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-3-662-45854-9 / DOI 10.1007/978-3-662-45854-9 (
#自动驾驶 #激光雷达 #视觉
什么是“BEV+Transformer”感知架构，它如何解决多摄像头融合的难题？
What is the BEV+Transformer perception architecture, and how does it solve multi-camera fusion?
BEV（鸟瞰图）把所有摄像头画面统一投影到一张俯视地图里，Transformer 则像注意力机制把不同视角的特征对齐融合。这样车不再“各看各的”，而是得到一张连贯的周围环视图。
BEV (bird's-eye view) projects all camera images into one top-down map, and a Transformer aligns and fuses features across views via attention. The car no longer sees separately but gets one coherent surround view.
难度 专业 · 适合 大众
CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source
出处备注：本题原挂出处与答案内容不符，已撤回；具体书目待补，暂不展示来源。
本题原挂出处与答案内容不符，已撤回；具体书目待补。
Previous citation withdrawn as unsupported for this answer; a specific source is pending.
#BEV #Transformer #多相机融合
无人系统（无人机/无人船/AUV）在失去通信时如何靠“自主导航”继续执行任务？
When unmanned systems (drones, unmanned ships, AUVs) lose communication, how do they keep missions going with autonomous navigation?
它们依靠惯性导航、地形或地磁匹配、以及板载任务重规划，在断联时按预设目标自行避障和决策。AUV 还能靠多普勒测速和声呐地图在水下无GPS环境定位。
They rely on inertial navigation, terrain or geomagnetic matching, and on-board mission replanning to avoid obstacles and decide on their own. AUVs additionally use Doppler velocity and sonar maps to navigate underwater without GPS.
难度 专业 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Thrun、Burgard & Fox，《概率机器人学》（2005）
Thrun, Burgard & Fox, 'Probabilistic Robotics' (2005)
独立核验依据 · Independent verification: 核验依据 1 ISBN 9780262332750 (2005)
#无人系统 #自主导航 #任务重规划
车路协同（V2X）是什么，为什么让“聪明车”加上“聪明路”更安全？
What is vehicle-to-everything (V2X) coordination, and why does pairing smart cars with smart roads improve safety?
V2X 让车与红绿灯、路侧设备和云端互相通信，能提前知道弯道后堵车或盲区里有行人，突破视距限制。车和路一起做决策，比单车闭眼硬扛更可靠。
V2X lets cars talk to signals, roadside units and the cloud, learning of hidden congestion or pedestrians beyond line of sight. Car and road decide together, more reliable than a lone car coping blindly.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Maurer、Gerdes、Lenz & Winner，《自动驾驶：技术、法律与社会视角》（Springer, 2016）
Maurer, Gerdes, Lenz & Winner, 'Autonomous Driving: Technical, Legal and Social Aspects' (Springer, 2016)
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-3-662-45854-9 / DOI 10.1007/978-3-662-45854-9 (
#V2X #车路协同 #超视距
自动驾驶汽车怎么知道“前面那辆车会不会突然并线”，决策规划靠什么做判断？
How does a self-driving car guess whether the car ahead will suddenly change lanes, and what does its planning rely on?
它先预测周围每辆车未来几秒的可能轨迹，再结合博弈和概率评估风险，最后在约束内选出最安全的加速度和转向。预测越准，紧急避险就越从容。
It predicts each nearby vehicle's possible future trajectories, assesses risk with game-theoretic and probabilistic reasoning, then picks the safest acceleration and steering within constraints. Better prediction means calmer emergency handling.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Maurer、Gerdes、Lenz & Winner，《自动驾驶：技术、法律与社会视角》（Springer, 2016）
Maurer, Gerdes, Lenz & Winner, 'Autonomous Driving: Technical, Legal and Social Aspects' (Springer, 2016)
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-3-662-45854-9 / DOI 10.1007/978-3-662-45854-9 (
#轨迹预测 #决策规划 #博弈
无人驾驶汽车靠什么“看”红绿灯和行人，它会不会闭着眼睛开车？
How does a driverless car 'see' traffic lights and pedestrians—does it ever drive with its eyes shut?
它用摄像头、雷达和激光像眼睛一样不停扫描周围，电脑马上认出红绿灯和走路的人。它一刻不停地“看”，比人还不敢闭眼。
It scans around continuously with cameras, radar and lasers, and the computer instantly recognizes lights and walking people. It 'watches' nonstop and dares not shut its eyes like we might.
难度 入门 · 适合 幼儿(3-6)
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Maurer、Gerdes、Lenz & Winner，《自动驾驶：技术、法律与社会视角》（Springer, 2016）
Maurer, Gerdes, Lenz & Winner, 'Autonomous Driving: Technical, Legal and Social Aspects' (Springer, 2016)
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-3-662-45854-9 / DOI 10.1007/978-3-662-45854-9 (
#无人驾驶 #感知 #传感器