Several innovations have combined to make that possible. Robotaxis rely on sensors such as cameras, laser-based lidars, microphones and radars to assess road conditions, judge distances and manage speed. They then use artificial intelligence both in the car and in the cloud to mimic the way human drivers process such information and react to it. And as robotaxi services have expanded, they have hoovered up ever more data, allowing them to refine their algorithms. The rise of multi-modal generative-AI models, which weave together text, images and sounds, has also accelerated progress, including by making it easier to train autonomous systems using simulations and teach them to react to unusual situations.
重点词汇:
innovation /ˌɪnə'veɪʃn/ n. 改革, 革新
combine /kəm'baɪn/ v.联合,结合
microphone /'maɪkrəfəʊn/ n. 扩音器, 麦克风
judge /dʒʌdʒ/ v.审判;评判
mimic /'mɪmɪk/ v.模仿,模拟
refine /rɪ'faɪn/ v.精炼,提纯,改善
algorithm /'ælɡərɪðəm/ n. 运算法则
accelerate /ək'seləreɪt/ v. (使)加快, (使)增速
simulation /ˌsɪmju'leɪʃn/ n. 模仿;模拟
几项创新结合起来使这成为可能。自动驾驶出租车依靠摄像头、激光雷达、麦克风和雷达等传感器来评估路况、判断距离和控制速度。然后,他们在汽车和云端使用人工智能来模仿人类驾驶员处理这些信息并做出反应的方式。随着自动驾驶出租车服务的扩展,他们收集了越来越多的数据,使他们能够完善自己的算法。将文本、图像和声音编织在一起的多模态生成人工智能模型的兴起也加速了进步,包括使使用模拟训练自主系统变得更容易,并教会它们对异常情况做出反应。