Yi Xiao, Felipe Codevilla, Akhil Gurram, Onay Urfalioglu, Antonio M. López
TL;DR本文提出并分析了终端到终端驾驶中,结合 RGB 和深度信息数据的多模态方案,以期提高自动驾驶 AI 模型的性能。通过模拟和条件模仿学习,证明了采用早期融合多模态方案,可以超越采用单模态的性能表现。
Abstract
A crucial component of an autonomous vehicle (AV) is the artificial
intelligence (AI) is able to drive towards a desired destination. Today, there
are different paradigms addressing the development of AI drivers. On the one
hand, we find modular pipelines, which divide the driving task